Universität zu Köln

Bomhard, Philipp von (2017). Development of a statistical daily precipitation model and its application to precipitation records. Masters thesis, Universität zu Köln.

Preview

PDF Masterarbeit_final.pdf - Accepted Version Download (11MB) | Preview

Abstract

In recent years, numerous record-breaking precipitation events have caused several deaths and high economic losses. However, current precipitation models do not adequately capture exceptionally high precipitation events. To address this problem, a new statistical daily precipitation model including several new aspects was developed. The model distinguishes between stratiform and convective precipitation, whereby the disaggregation of these two types of precipitation is based on SYNOP reports of 300 German weather stations. By combining a Weibull distribution with power law, a new probability distribution was derived and later implemented in the amount process. This four-parameter distribution improves the modelling of extreme precipitation amounts enormously and still delivers good results for small amounts. To take account of variability, a confined random walk was additionally implemented in the occurrence process of the model. The model was developed in a manner that allows universal use. By incorporating the elevation of a station and the time of the year as input parameters, the model was made applicable at any time and for any location in Germany. In a second step the model was applied to investigate daily precipitation records. For that, linear changes were implemented into the model. As result, a previously found decrease of records in the summer season can be explained by changes in the stratiform precipitation distribution. However, the decrease of precipitation records in the summer season is too low to rule out random processes as cause of this decrease. In contrast, the increase in the mean number of precipitation records in winter season cannot be reproduced with the developed model. A possible explanation for that is the neglect of spatial correlations in the amount process. An appropriate method for taking spatial correlations into account, could be a Copula approach.

Item Type:	Thesis (Masters thesis)
Translated title:	Title Language Entwicklung eines statistischen Niederschlagmodells und dessen Anwendung zur Untersuchung von Niederschlagsrekorden German
Creators:	Creators Email ORCID ORCID Put Code Bomhard, Philipp von philipp.vonbomhard@deutscherueck.de UNSPECIFIED UNSPECIFIED
URN:	urn:nbn:de:hbz:38-91859
Date:	8 September 2017
Place of Publication:	Köln
Language:	English
Faculty:	Faculty of Mathematics and Natural Sciences
Divisions:	Faculty of Mathematics and Natural Sciences > Department of Geosciences > Institute for Geophysics and Meteorology Faculty of Mathematics and Natural Sciences > Department of Physics > Institute for Theoretical Physics
Subjects:	Physics Earth sciences
Uncontrolled Keywords:	Keywords Language precipitation model records record statistk English Niederschläge Modell Niederschlagsmodell Rekorde Rekordstatistik German
Date of oral exam:	19 October 2017
Referee:	Name Academic Title Krug, Joachim Prof. Dr. Shao, Yaping Prof. Dr.
Refereed:	Yes
URI:	http://kups.ub.uni-koeln.de/id/eprint/9185