Continuous and discrete wavelet transforms based analysis of weather data of North Western Region of Saudi Arabia
| Pages | 369-389 |
| Date | 01 November 2010 |
| Published date | 01 November 2010 |
| DOI | https://doi.org/10.1108/20425945201000023 |
| Author | Mohamed A. El‐Gebeily,Shafiqur Rehman,Luai M. Al‐Hadhrami,Jaafar AlMutawa |
| Subject Matter | Public policy & environmental management |
World Journal of Science, Technology and Susainable Development, Vol. 7, No. 4, 2010
369
Copyright © 2010 WASD
Abstract: The present study utilizes daily mean time series of meteorological
parameters (air temperature, relative humidity, barometric pressure and wind
speed) and daily totals of rainfall data to understand the changes in these
parameters during 17 years period i.e. 1990 to 2006. The analysis of the above
data is made using continuous and discrete wavelet transforms because it provides
a time–frequency representation of an analyzed signal in the time domain.
Moreover, in the recent years, wavelet methods have become useful and powerful
tools for analysis of the variations, periodicities, trends in time series in general
and meteorological parameters in particular. In present study, both continues
and discrete wavelet transforms were used and found to be capable of showing
the increasing or decreasing trends of the meterorological parameters with.
The seasonal variability was also very well represented by the wavelet analysis
used in this study. High levels of compressions were obtained retaining the
originality of the signals.
Keywords: weather; meteorology; compression, decomposition, wavelet transform, trend
analysis
Mohamed A. El-Gebeily1, Shaqur Rehman2 ,
Luai M. Al-Hadhrami2 and Jaafar AlMutawa1
King Fahd University of Petroleum and Minerals, Saudi Arabia
Continuous and disCrete Wavelet
transforms Based analysis of
Weather data of north Western
region of saudi araBia
introduCtion
Climate change on global, regional, and lo-
cal scales is of great concern and has been
the focus of attention of many researchers
in the fields of science, engineering and
social studies throughout the world. This
is because the long-term climate variabil-
ity is of great importance for the estima-
tion of its impact on human activities and
for predicting future behavior. Over the
past century or so the world has warmed
by approximately 0.6°C, as quoted by
Nicholas and Collins (2006). According
to IPCC report (2001) and Meehl et
al, (2004), there is strong evidence that
most of the global warming over the past
50 years is likely to have been due to in-
creases in greenhouse gas concentrations.
The climatic studies are very common
1 Department of Mathematics and Statistics, King Fahd University of Petroleum and Minerals, Dhahran-31261,
Saudi Arabia, E-mail: mgebeily@kfupm.edu.sa, jaafarm@kfupm.edu.sa, Homepage: http://faculty.kfupm.edu.sa/
math/mgebeily http://faculty.kfupm.edu.sa/math/jaafarm
2 Center for Engineering Research, Research Institute, King Fahd University of Petroleum and Minerals,
Dhahran-31261, Saudi Arabia, E-mail: srehman@kfupm.edu.sa Homepage: http://faculty.kfupm.edu.sa/ri/srehman
370 M. A. El-Gebeily, S. Rehman, L. M. Al-Hadhrami and J. AlMutawa
and have been conducted in almost every
part of the world, for example Freiwana
and Kadioglub (2008) for Jordan, Elagib
and Addin Abdu (1997) for Bahrain and
Alkolibi (2002) for Saudi Arabia.
The wavelet transform is a strong mathe-
matical tool that provides a time–frequency
representation of an analyzed signal in the
time domain as reported by Dabuechies
(1990) and Percival and Walden (2000).
Currently, wavelet methods are being used
as powerful tools for the analysis of varia-
tions, periodicities, trends in time series
(Partal and Kucuk, 2006; Pisoft et. al., 2004;
and Yueqing et al., 2004).
Wavelet transforms have been used to
investigate trends in the central England
temperature series (Baliunas et al., 1997); to
study hemispheric temperature series and
the southern oscillation index (Sonechkin
et al., 1999); to detect shifts in global tem-
perature (Park and Mann, 2000); and to an-
alyze variability in European temperatures
(Datsenko et al., 2001). Ding et. al. (2002)
used wavelet transform to understand the
frequency features of the Hong Kong tem-
perature data. The climate parameters were
analyzed by Lau and Weng (1995) using
wavelet methods to detect and highlight the
climatic features of the signal.
In this work we use continuous and dis-
crete wavelet transforms to analyze the mete-
orological records (temperature, barometric
pressure, precipitation, wind speed and rela-
tive humidity) of the weather station at Arar
in the north western part of Saudi Arabia
over the 17 year period from 1990 to 2006.
Our goals are to identify the long term trends,
detect periodicity and anomalous events and
to study the compression of the records. It
will be seen that wavelet tools provide clear
indications of the sought events as well as a
powerful tool for data compression.
site and data desCription
Arar is a town located in the north western
part of Saudi Arabia. The latitude, longi-
tude and the altitude of the data collection
station are 30°54’, 41°08’ and 542 meters,
respectively. The population of this region
has increased during last couple of decades
and economical growth has taken place. As
a result of which the number cars on the
road, air transport in the region, the infra-
structure, the support services, etc. have
increased exponentially. So, to understand
the local effect on weather parameters of
the region, daily mean values of air tem-
perature, relative humidity, barometric pres-
sure, wind speed and daily totals of rainfall
data are used during the period of 1990
to 2006. This data is collected and man-
aged by the Presidency of Meteorology and
Environment (PME) at the national airport
in Arar.
The air temperature at Arar was found
to vary between a minimum of -0.7°C and
40.4°C while the overall mean remained
as 22.2°C. In this region, higher air tem-
peratures of >25°C were observed during
the months of May to September during
the year. The minimum temperatures were
observed in the months of December and
January. The barometric pressure varied
between a minimum of 936mb and a maxi-
mum of 967mb while the overall mean re-
mained as 949.6mb. The relative humidity
was found to vary between a minimum of
17.7% and a maximum of 65%. The overall
mean relative humidity in the region dur-
ing entire data collection period was 36%.
Higher values of relative humidity were ob-
served during winter months (October to
March) and lower during May to September.
The wind speed was found to vary between
0 and 25 knots while the overall mean was
found to be 7 knots. Higher wind speeds
were observed in the months of January
To continue reading
Request your trial