Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Statistical Analysis of Oceanographic Data: A Comparison between Stationary and Mobile Sea Level Gauges

Posada, Marcus (2014) FMS820 20141
Mathematical Statistics
Abstract (Swedish)
This Master’s thesis project concerns developing statistical methods to examine
sea level data provided by the Swedish Meteorological and Hydrological
Institute (SMHI). The data comes from sea level gauges located in the harbours
of Uddevalla, Ängelholm and Åhus along the Swedish coast. These
three gauges are mobile, as compared to the permanent sea level gauges
used by the SMHI at various points along the coast. The mobile gauges
were placed during 2010 and the goal of this project is to analyse the data
they have provided. This is done in several steps as outlined below.
Initially, a comparison is performed between the extreme sea levels at
the three mobile gauge locations and permanent gauges placed nearby. This
analysis... (More)
This Master’s thesis project concerns developing statistical methods to examine
sea level data provided by the Swedish Meteorological and Hydrological
Institute (SMHI). The data comes from sea level gauges located in the harbours
of Uddevalla, Ängelholm and Åhus along the Swedish coast. These
three gauges are mobile, as compared to the permanent sea level gauges
used by the SMHI at various points along the coast. The mobile gauges
were placed during 2010 and the goal of this project is to analyse the data
they have provided. This is done in several steps as outlined below.
Initially, a comparison is performed between the extreme sea levels at
the three mobile gauge locations and permanent gauges placed nearby. This
analysis results in estimates of five-year return levels for the differences between
the mobile and nearby permanent gauges. It turns out that the difference
is roughly 50 cm. The second analysis involves studying the relationship
between the sea levels at the paired stations using linear regression. As a final
analysis these models are diagnosed and extended to include parameters
for wind speed, wind direction and atmospheric pressure.
The largest single objection to the validity of the regression models is
arguably the natural temporal dependence in the data which indicates that
time series models could be more suitable. Such models are fitted to the
dataset where fractional integration is applied to handle long-term memory
and GARCH models are implemented for conditional heteroscedasticity. In
addition, harmonic tidal analysis is also performed. (Less)
Please use this url to cite or link to this publication:
author
Posada, Marcus
supervisor
organization
course
FMS820 20141
year
type
H2 - Master's Degree (Two Years)
subject
language
English
id
4362127
date added to LUP
2014-03-24 12:17:06
date last changed
2014-03-24 12:17:06
@misc{4362127,
  abstract     = {{This Master’s thesis project concerns developing statistical methods to examine
sea level data provided by the Swedish Meteorological and Hydrological
Institute (SMHI). The data comes from sea level gauges located in the harbours
of Uddevalla, Ängelholm and Åhus along the Swedish coast. These
three gauges are mobile, as compared to the permanent sea level gauges
used by the SMHI at various points along the coast. The mobile gauges
were placed during 2010 and the goal of this project is to analyse the data
they have provided. This is done in several steps as outlined below.
Initially, a comparison is performed between the extreme sea levels at
the three mobile gauge locations and permanent gauges placed nearby. This
analysis results in estimates of five-year return levels for the differences between
the mobile and nearby permanent gauges. It turns out that the difference
is roughly 50 cm. The second analysis involves studying the relationship
between the sea levels at the paired stations using linear regression. As a final
analysis these models are diagnosed and extended to include parameters
for wind speed, wind direction and atmospheric pressure.
The largest single objection to the validity of the regression models is
arguably the natural temporal dependence in the data which indicates that
time series models could be more suitable. Such models are fitted to the
dataset where fractional integration is applied to handle long-term memory
and GARCH models are implemented for conditional heteroscedasticity. In
addition, harmonic tidal analysis is also performed.}},
  author       = {{Posada, Marcus}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Statistical Analysis of Oceanographic Data: A Comparison between Stationary and Mobile Sea Level Gauges}},
  year         = {{2014}},
}