Sunteți pe pagina 1din 92

Investigating the Galaxy Distribution using

Redshift Surveys

Sara Nabhan Al-Battashi

A thesis submitted on partial fulfillment of the requirements for


the degree of
Master of Sicence
in Physics

Department of Physics
College of Science
Sultan Qaboos University
Sultanate of Oman
(October, 2014)

Thesis Committee

Thesis Examining Committee

II

Acknowledgment

Firstly and above all else, praise and thanks to Allah, the almighty for His
guidance throughout my research and for granting me the capability to proceed
successfully.
My thanks go to everyone who helped me in the completion of this project,
beginning with my thesis supervisor, Dr. Saleh Al-Shidhani, for initially introducing
me to the field of cosmology, suggesting it as a topic for my project, responding to my
questions and queries in such a prompt manner and finally his patience in resolving the
technical problems that I faced along the way. Without his help I would not have been
able to complete the project in the amount of time that I did.
I would also like to express my deep and sincere gratitude to Dr. Laila Alabidi
for the fruitful discussions that I had with her, as well as for her assistance with the
statistical techniques used in the project. Many thanks to Dr. Randa Asaad for her
sheer interest in the subject matter of this research, and also for the valuable time that
she put into our insightful discussions about this project.
I would also like to thank my examiners, Dr. Milan Bogosavljevic and Dr.
Zacharias Ioannou for the time they gave up to read my thesis and giving the valuable
suggestions and corrections to this work.
My sincere thanks also goes to Prof. Matthew Colless for his help with the
transformations of the coordinates.
I submit my highest appreciation to my academic advisors, Prof. Abraham
George and Dr. Salim Al-Harthi, for their valuable advice and their help along every
step of the way.
I am extremely grateful to my family, particularly to my mother for her prayers
and her caring, along with the sacrifices that she has made to educate and prepare me for
my future. A warm thanks to my brothers and sisters and their families for their
encouragement and support in so many ways.
I also wish to extend my thanks to Oliver Allan for the linguistic support during
the writing process and over the past two years, and Muna Al-Sawafi, for allowing me
to learn from her experience.
III

Abstract

Understanding the structure and evolution of the universe has been and remains a
very interesting area of research. Many researchers carried out several analyses with
different sky survey sizes. In this study, a large sample of galaxies and QSQs has been
constructed from ten redshift sky surveys to study the distribution of the galaxies
throughout universe. The sample covers a large area of the sky and comprises of
2763064 objects, mostly obtained from the SDSS Survey. The study is focused on the
radial distribution of celestial objects, specifically in testing the claim of periodicity
(also referred to as the quantization) in the galaxy redshifts. The phenomenon of
periodicity was investigated using an unbiased sample of high quality redshift
measurements and with periodogram spectral estimation. The radial distribution in the
comoving scale was found to exhibit a periodic separation of about ~167 Mpc. No
evidence for a periodicity has been found in the redshift (z) scale.

IV


.
.

. 2763064
) .(SDSS
( ) .

periodogram .
~167 .
.

Table of Contents
Chapter 1: Introduction ...................................................................................................... 1
Chapter 2: The Expanding Universe and the Large-Scale Structure Formation ................ 3
2.1. History of the Big Bang Theory .............................................................................. 3
2.2. Large-Scale Structure Formation and Evolution ..................................................... 5
2.3. Structure Periodicity ................................................................................................ 7
Chapter 3: Redshift and the Hubble Law ......................................................................... 10
3.1. Redshift Effect ....................................................................................................... 10
3.2. Redshift Measurements ......................................................................................... 12
3.3. Discovery of the Hubbles Law ............................................................................. 13
3.4. Hubbles Constant ................................................................................................. 15
Chapter 4: The Redshift Surveys, Past, Present and Future ............................................. 18
4.1. The Survey Methods ............................................................................................. 18
4.1.1. Pencil-beam Surveys. .................................................................................. 18
4.1.2. Slice Surveys. .............................................................................................. 19
4.1.3. All-Sky and Large Surveys. ............................................................................ 19
4.1.4. Targeted Surveys with special kinds of objects. . .......................................... 19
4.1.5. Blind Redshift Surveys carried out in the 21 cm line of natural hydrogen
(HI)............................................................................................................................ 20
4.2. Sloan Digital Sky Survey (SDSS) ......................................................................... 20
4.3. Two-degree-Field Galaxy Redshift Survey (2dFGRS) ......................................... 21
4.4. Six-degree-Field Galaxy Redshift Survey (6dFGRS) ........................................... 22
4.5. VIMOS-VLT Deep Survey (VVDS) ..................................................................... 23
4.6. The Deep Extragalactic Evolutionary Probe project (DEEP) ............................... 23
4.7. Two Micron All-Sky Survey (MASS) .................................................................. 24
4.8. VIMOS Public Extragalactic Redshift Survey (VIPERS)..................................... 24
4.9. Galaxy And Mass Assembly Survey (GAMA) ..................................................... 24
4.10. FORS Deep Field (FDF) spectroscopic survey ................................................... 25
4.11. Team Keck Redshift Survey (TKRS) .................................................................. 25
Chapter 5: Data Collection and Statistical analysis ......................................................... 26
5.1 Checking data consistency ..................................................................................... 26
VI

5.1.1 Coordinate Transformations. ........................................................................... 27


5.1.2 Duplication checking ....................................................................................... 29
5.2 Checking for periodicity ......................................................................................... 30
5.2.1 Histogram......................................................................................................... 31
5.2.2 Periodogram ..................................................................................................... 31
Chapter 6: Results and discussion .................................................................................... 33
6.1 Radial Distributions ................................................................................................ 34
6.1.1 Histograms. ...................................................................................................... 34
6.1.2 Periodograms ................................................................................................... 53
6.2. Visual explorations of clustering and connectivity ............................................... 64
Chapter 7: Summary and Conclusion .............................................................................. 72
Appendix A: ..................................................................................................................... 73
A MATLAB program for the periodogram calculations ................................................. 73
References ........................................................................................................................ 75

VII

List of Tables
Table 6.1: Summery of the redshift measurements before and after filtering process. 33

VIII

List of Figures
Figure 2.1: Computer simulation of structure formation. Each side of the boxes are 43
Mpc wide. The left-most frame is simulation of the universe when it was less than 1%
of its current age (z=30) and the right-most frame is the illustration of the present day
universe. 6
Figure 2.2: Galactocentric differential redshifts of the 48 Virgo spirals, in bins 11 km s-1
wide. No data smoothing has been applied. Dotted vertical lines represent periodicity
71.1 km s-1 and zero phase. .. 7
Figure 2.3: Galactocentric periodicity of ~ 37.5 km s-1 observed in the differential
redshifts of 97 bright spiral galaxies scattered throught the Local Supercluster. . 8
Figure 3.1: The original Hubble diagram of 1929. The plot gives the observed redshift
(again expressed as a Doppler shift given in km s-1) as a function of distance in (parsec)
. 15
Figure 3.2: Cosmic Microwave Background seen by Planck. ...... 17
Figure 4.1: The first set of observations done for the CFA redshift survey in 1985 by
Valerie de Lapparent, Margaret Geller and John Huchra. .. 19
Figure 4.2: The distribution of 63000 2dFGRS galaxies in the NGP (left panel) and SGP
(right panel) strips. .. 22
Figure 5.1: The commoving distance as a function of z, based on equation 5.1. .. 28
Figure 5.2: The difference between the redshifts with respect to the equatorial and
galactic coordinates. ... 30
Figure 6.1: The collective sample redshifts represented using histograms with four
different bin sizes: 0.2, 0.1, 0.01 and 0.001 (From up to bottom). . 36
Figure 6.2: Histogram of the collective sample redshifts with bin size of 0.01. ... 37
Figure 6.3: Histogram of the collective sample redshifts with bin size of 0.01. The
Frequency is represented using the logarithmic scale. ... 38
Figure 6.4: Histogram of the SDSS sample redshifts with bin size of 0.01. . 39
IX

Figure 6.5: Histogram of the SDSS sample redshifts with bin size of 0.01.

The

Frequency is represented using the logarithmic scale. ... 40


Figure 6.6: Histogram of the 2MASS sample redshifts with bin size of 0.01. .. 41
Figure 6.7: Histogram of the 2MASS sample redshifts with bin size of 0.01. The
Frequency is represented using the logarithmic scale. ... 42
Figure 6.8: Histogram of the 2dF sample redshifts with bin size of 0.01. 43
Figure 6.9: Histogram of the 2dF sample redshifts with bin size of 0.01. The Frequency
is represented using the logarithmic scale. . 44
Figure 6.10: Histogram of the 6dF sample redshifts with bin size of 0.01. .. 45
Figure 6.11: Histogram of the 6dF sample redshifts with bin size of 0.01.

The

Frequency is represented using the logarithmic scale. ... 46


Figure 6.12: Histogram of the GAMA sample redshifts with bin size of 0.01. 47
Figure 6.13: Histogram of the VIPERS sample redshifts with bin size of 0.01. ... 48
Figure 6.14: Histogram of the DEEP sample redshifts with bin size of 0.01. ...... 49
Figure 6.15: Histogram of the VVDS sample redshifts with bin size of 0.01. .. 50
Figure 6.16: Histogram of the TKRS sample redshifts with bin size of 0.01. .. 51
Figure 6.17: Histogram of the FDF sample redshifts with bin size of 0.01. . 52
Figure 6.18: Periodograms of the collective sample redshifts calculated using four
different sampling rates. . 56
Figure 6.19: The corresponding cumulative periodograms of Figure 6.18s
periodograms. . 57
Figure 6.20: Periodograms of the redshifts of sample 2 (the well-sampled region)
calculated using four different sampling rates. ... 58
Figure 6.21: The corresponding cumulative periodograms of Figure 6.20s
periodograms. . 59

Figure 6.22: Periodograms of sample 1 (upper panel) and sample 2 (lower panel)
represented using the logarithmic scale. ..... 60
Figure 6.23: Periodogram of sample 1 for the high frequency range. ... 61
Figure 6.24: Periodogram of sample 2 for the high frequency range. ... 62
Figure 6.25: The periodogram of sample 3s data. 63
Figure 6.26: The cumulative periodograms of sample 3s data. ... 64
Figure 6.27: Periodogram of sample 3 represented using the logarithmic scale.
..... 65
Figure 6.28: Periodogram of sample 3 for the high frequency range. ... 66
Figure 6.29: The 3-D distribution of galaxies covered by the sample, represented in the
Cartesian coordinates converted from equatorial coordinates. ... 68
Figure 6.30: The 3-D distribution of galaxies covered by the sample, represented in the
Cartesian coordinates converted from galactic coordinates ... 69
Figure 6.31: The 2-D distribution of galaxies covered by the sample along the
equatorial coordinates; ra and dec. ..... 70
Figure 6.32: The 2-D distribution of galaxies covered by the sample along the galactic
coordinates; l and b. .... 71

XI

Chapter 1: Introduction
The curiosity about the structure, history and state of the universe has always
encouraged the scientists to investigate the galaxy distribution in the universe, as it is the
key to understand it. However, the distance estimation of astronomical objects was
always the prime challenge that limited our ability to map the structure. Throughout
history, astronomers developed various techniques to measure distances to the
astronomical objects at different ranges, starting from the radar ranging technique which
is limited to the solar system studies and ending with the application of the Hubbles law
(or the redshift phenomenon) that allows measuring the distance to very far extragalactic
objects.
By the 1970s, projects called Redshift Sky Surveys were established to
measure the distances of a large number of astronomical objects, based on Hubbles law
and the redshift phenomenon. The survey catalogues provide the redshifts of the
astronomical objects (usually galaxies) combined with their angular positions and some
photometric properties.

Therefore, they can be used to construct a comprehensive

picture of the universe. From such picture, the aim was to examine the Large Scale
Structure (LSS) distribution and features as well as to inform the modeling of the LSS
evolution.
Many studies (e.g. (Hawkins, Maddox, & Merrifield, 2002) and (Tang & Zhang,
2005)) used various methods to quantify the LSS of the universe and trace its evolution
and origin, such as the implementation of the statistical analysis on the data provided by
the Redshift Sky Surveys to examine specific features of the spatial distribution of the
astronomical objects like the degree of clustering and connectivity, looking for the
presence of voids and the detection of periodic structures. In many cases, the sample
selection procedures and the statistical effects can significantly bias the results of the
study and lead to draw unreliable conclusions.

Therefore, the selection of a

representative sample of the universe and the application of the appropriate statistical
technique are critical for providing accurate results that can be generalized to the whole
universe.
In this study, we used a collection of ten publicly available Redshift Survey data
releases, namely: Sloan Digital Sky Survey (SDSS, 10th Release), Two Micron All-Sky
1

Survey (2MASS, the final data product), Two-degree-Field Galaxy Redshift Survey
(2dFGRS, the best spectroscopic observations of the final release), Six-degree-Field
Galaxy Redshift Survey (6dFGRS, Data Release 3), Galaxy And Mass Assembly survey
(GAMA, Data Release 2), VIMOS Public Extragalactic Redshift Survey (VIPERS,
PDR-1), The Deep Extragalactic Evolutionary Probe project (DEEP, Data Release 4),
VIMOS-VLT Deep Survey (VVDS, First Epoch sample), Team Keck Redshift Survey
(TKRS) and FORS Deep Field spectroscopic Survey (FDF). We used this collection to
examine a controversial issue regarding the structure of the universe: The claim of the
periodic distribution of galaxies in space, which is known as the redshift periodicity
phenomenon. Most of the previous studies on the phenomenon (e.g. (Napier & Guthrie,
1997)) used relatively small and biased samples, however with this project we aim to
test the hypothesis using a large data set and by using suitable statistical techniques.
After this introductory chapter, the dissertation will briefly discuss the
development of the Big Bang Theory from a historical angle as well as the evolution of
the Large Scale Structure of the universe in chapter 2. The last section of chapter 2
(section 2.3) will introduce the problem that we have studied in this project: the Redshift
Periodicity, as well as the previous researches that pertain to it. In chapter 3, the redshift
phenomenon and the discovery of the Hubbles Law will be explained in some detail.
Chapter 4 is dedicated to the Redshift Surveys: their development, the types of
Redshift Surveys and details about the ten redshifts surveys which have been used in this
project. The methodology that is used in this research to collect the data, ensure its
reliability and test the redshift periodicity will be provided in chapter 5. The results of
the work and the discussion will be presented in chapter 6. Finally, chapter 7 will
contain a summary of the work that has been carried out, the conclusion we have
reached and suggestions for further work.

Chapter 2: The Expanding Universe and the Large-Scale Structure Formation


For thousands of years, astronomers wondered about the universe, questioning a
number of things regarding its age, size, structure and our place within it. Moreover,
they were curious to know how it began (if it has a beginning) and how matter came to
exist. Throughout history, there have been many attempts to answer these questions, but
before the twentieth century, most of the attempts failed to yield an integrated idea about
the universe that gives consistent answers to all of the questions. At the beginning of the
twentieth century, several observations supported by theoretical considerations led to the
establishment of todays best explanation for how the universe began; the Big Bang
theory. According to this theory the universe began at a moment in the distant past and
evolved from a very hot, dense and almost uniform structure into the complicated
system that we see today.
In this chapter, a brief history of the Big Bang theory and the evolution of the
Large-Scale structure according to it will be presented in the first section. The next
section will focus on an issue concerning the distribution of matter in space: The redshift
quantization, also known as redshift periodicity.
2.1. History of the Big Bang Theory
The seed for the idea of the Big Bang came from Albert Einstein in his field
equations of general theory of relativity in the early twentieth century, but Einstein
himself didnt like it because he was convinced; as was everybody at that time, with the
model of a static and eternal universe. In general relativity, he proposed that mass warp
space and time to create gravity, but if gravity is always pulling in, then what keeps
everything from ultimately fusing together into one massive object. Einstein believed
that there must be another equal counter force pushing out in opposition to gravity,
keeping the eternal universe in perfect balance. Therefore, in 1917 he inserted a positive
cosmological constant into his general theory of relativity to force the equations to
predict a stationary universe and match the observations of that time that strongly
favoured a steady universe (Straumann, 2002). A few months later in the same year,
1917, Willem de Sitter proposed another solution to the field equations that produced a
non-expanding, static universe if it contains no matter. In contrast to Einsteins universe

that contained matter but no motion, de Sitters universe involved motion without matter
(Encyclopedia of Time, 1994).
In 1922, a Russian mathematician called Alexander Friedmann derived a solution
to the general relativity field equations that described mathematically a model of an
expanding universe, before there was any observational evidence (Tropp, Frenkel, &
Chernin, 2006). In 1924 he published a paper that described a homogenous, isotropic
and expanding or contracting universe; the Friedmann-Lemare-Robertson-Walker
universe, which is named after the four astronomers: Alexander Friedmann, George
Lemare, Howard P. Robertson and Arthur Geoffrey Walker, who developed the model
independently (Godart & Heller, 1985).
Five years later, Lemare published his Ph.D. research in which he concluded
that the universe was expanding based on Einsteins general theory of relativity and the
redshift measurements of the extragalactic nebulae (L'Annunziatea, 2007).
At the observational level, the most revolutionary step was the discovery of the
linear relationship between the rate of recession of distant galaxies (calculated from their
observed redshifts) with the distance to them, which was made by Edwin Hubble. The
relationship, which is now known as Hubbles Law, was first published by Hubble in
1929 and then refined and published later in association with Milton Lasell Humason, in
1931 (Belenkiy, 2014). The law implies that farther galaxies are going away from us at
higher speeds; hence their spectra are shifted from shorter wavelengths to longer
wavelengths as the light travels from the galaxy to us and the amount of shift is
proportional to distances. This increase in the wavelength is attributed to the expansion
of the space and led to a model of the universe that is consistent with the solutions of
Einsteins General Relativity Equations for a homogenous, isotropic expanding space.
Hubbles observations encouraged other cosmologists such as Arthur Eddington,
de sitter and Einstein to realise that the static models of the universe were unsatisfactory
and to become aware of Lemaitres 1927 paper (Soter & Tyson, 2000).
In 1931, Lemaitre published several papers in which he formulated
mathematically an expanding and homogenous model of the universe and proposed that
it began with an enormous explosion that would start the expansion of the universe

(Angelo, 2009). The explosion which Lemaitre referred to is what came to be known
as the Big Bang.
From its discovery until the present, the Big Bang is the most popular and
acceptable model of the formation of the universe.
2.2. Large-Scale Structure Formation and Evolution
The Big Bang model assumes that the universe began 13.8 billion years ago
(Bennett, et al., 2013) when space and time were generated in a vast explosion known as
the Big Bang, creating all matter and energy in the known universe via a process called
inflation that lasts until t=10-35 seconds after the Big Bang (Ebrahimi & Riazi, 2012). In
the next one second, the inflation of the universe ended quickly and after the rapid initial
expansion, the universe began to slow down.

The four fundamental forces: the

gravitational force, the strong force, the weak force, and the electromagnetic force took
shape and began to hold the universe together . Then in the following three minutes, the
universe as we know today began to take shape; protons and neutrons came together to
form the basic matter and elements (mostly hydrogen and helium). In the next 500,000
years, the universe remained a huge cloud of expanding gas that eventually cooled
enough that celestial bodies could form. Photons from this period have remained as the
Cosmic Microwave Background radiation (CMB) and can still be observed today
(O'Callaghan, 2012).
Initially, the universe was in a very dense, hot, nearly uniform, and isotropically
expanding state (Relativity in General, 1994) , and as time went on, it cooled from 1032
to 2.73 kelvin (Dardo, 2004) and the density fluctuations grew until the universe
structure became the complicated system that we see today. A few hundred million
years after the Big Bang (Bodenheimer, 2011), some matter clumped together to form
stars. Then gravity played new role and galaxies began to form some three billion years
after the Big Bang (Woolfson, 2013). These galaxies interacted gravitationally and
hence grouped into clusters. Clusters can vary from consisting of just a few members,
as in the Local Group that our Milky Way galaxy is a part of (Typically, clusters with
fewer than 50 members are called groups rather than clusters), to clusters containing
thousands of galaxies which are the largest known gravitationally bound systems in the
universe. However, the physical size of all the clusters is typically of the same order,
5

regardless of the number of galaxies contained in the cluster, which means that richer
clusters (clusters with large number of galaxies) have higher densities. For example, the
typical cluster size of a few megaparsecs is not much different from the diameter of the
Local Group (Jones & Lambourne, 2004).
Clusters themselves combine to form larger structures called superclusters. In
contrast to the clusters of galaxies, superclusters are non-gravitationally bound and
consequently, they can get involved in the cosmic expansion and are slowly dispersed by
the expansion of the universe (Appenzeller, 2009) . Some of the superclusters have the
shape of a wall (e.g. Northern Great Wall) and some bear a shape closely resembling a
filament (eg. Sculptor Wall), and both surround empty spaces known as voids (Vicent J.
Martine & Enn Saar, 2002).
These stages of the universe evolution can be illustrated using a computer
simulation such as the one shown in Figure 2.1. The left-most frame represents the old
form of the universe when it was very dense and has less fluctuations compared to the
present day universe (right-most frame) where the clusters and filaments can be clearly
seen.

Figure 2.1: Computer simulation of structure formation. Each side of the boxes are 43 Mpc wide. The left-most
frame is simulation of the universe when it was less than 1% of its current age (z=30) and the right-most frame is
the illustration of the present day universe. (Cosmic Web and Formation of Galaxies ).

In the prior few paragraphs, we discussed the formation and distribution of the
LSS building blocks in a rather general sense; however in the next sections we are going
to discuss the distribution through a specific topic: The claims for redshift periodicity.

2.3. Structure Periodicity


Since the 1970s, claims that galaxies are found in regularly-spaced groups started
to appear by researchers such as William Tifft and his colleagues. They observed that
various astronomical objects tend to favour multiples of some particular values in their
velocities. Tifft, in his 1979 paper Periodicity in the Redshift Intervals for Double
Galaxies (Tifft, 1979) reported finding two possible periodicities:
71 km/s observed for the differential redshifts* between galaxies in
groups
36 km/s in the redshifts of galaxies measured with respect to our own
galactic centre
Several subsequent studies have confirmed Tiffts findings; such as (Napier &
Guthrie, 1997) where in the galactocentric differential redshifts of the 48 Virgo spirals
were found to be quantized in steps of 71.1 km/s and a galactocentric periodicity of
~37.5 km/s was observed in the differential redshifts of 97 bright spiral galaxies
scattered through the Local Supercluster. The two periodicities found by Napier and
Guthrie are displayed in Figures 2.2 and 2.3.

Figure 2.2: Galactocentric differential redshifts of the 48 Virgo spirals, in bins 11


-1
km s wide. No data smoothing has been applied. Dotted vertical lines
-1
represent periodicity 71.1 km s and zero phase. From (Napier & Guthrie, 1997)
*

The differential redshifts (obtained by subtracting redshifts in pairs) are used to discriminate the local
motions from the recessional velocity due to the cosmic expansion. However, this technique is valuable
only for objects that lie close together in the sky, as with pairs and groups, because when galaxies from
different regions of the sky are involved, different motions contribute differently to the redshifts. In the
latter case, the redshifts themselves must be used.

-1

Figure 2.3: Galactocentric periodicity of ~ 37.5 km s observed in the differential


redshifts of 97 bright spiral galaxies scattered throught the Local Supercluster.
(Napier & Guthrie, 1997).

Later studies indicated that velocity breaks could also occur at submultiples of 71
km/s, like 1/2, 1/3 and 1/6 (Feigelson & Babu, 1996) (Trimble, 2000).
However, even after all of this evidence for periodicity, there is a scientific resistance to
the idea. Many scientists criticize the redshift quantization and ascribe the phenomenon
to statistical artefact and selection procedures. They also point out that the results tend to
come from a small group of astronomers who have a strong prejudice in favour of
detecting such unconventional phenomenon (Hawkins, Maddox, & Merrifield, 2002).
Hawkins, Maddox and Merrifield used Napiers own guidelines for testing redshifts in
much larger sample and they found no evidence for a redshift periodicity and stated that
... The criticism usually leveled at this kind of study is that the samples of redshifts
tended to be rather small and selected in a heterogeneous manner, which makes it hard
8

to assess their significance (Hawkins, Maddox, & Merrifield, 2002). Also, Su Min
Tang and Shuang Nan Zhang tested the hypothesis in the light of two existing models,
namely the Karlsson log(1+z) model and Bells decreasing intrinsic (DIR) Model, using
the data of Sloan Digital Sky Survey and 2dF QSO redshift survey and the concluded
that there is no evidence for periodicity (Tang & Zhang, 2005).
However, the aim of this project is to examine the periodicity hypothesis without
lending any bias to a specific result, using a large data set and an appropriate statistical
analysis in order to avoid the defects of the previous similar studies.

Chapter 3: Redshift and the Hubble Law


In 1842 an Austrian mathematician and physicist called Christian Doppler
published a study on coloured light emanating from binary stars. Doppler suggested that
the colour of the light from a star is dependent on the stars velocity relative to Earth.
This theory became known as the Doppler Effect and is extensively used in astronomy
to study the motion of objects across the universe. In the early nineteenth century, the
American astronomer Vesto Slipher observed that the spectra of the vast majority of
distant galaxies are shifted to longer wavelengths, and relying on the Doppler Effect, he
concluded that they are receding. Later on, another American astronomer called Edwin
Hubble made a great discovery when he combined measurements of galaxy distances
with measurements of the redshifts associated with the galaxies, and found a rough
linear proportionality between the objects distance and its redshift. The distanceredshift relation which is now known as the Hubbles Law has become an important tool
for mapping the universe in three dimensions.
In this chapter, we present an overview of the redshift effect and its
measurements and how this lead us to imagine an expanding universe and the theory that
the universe began with the Big Bang 13.8 billion years ago.
3.1. Redshift Effect
The Doppler Effect states that the wavelength of waves changes, hence the
frequency of waves does too if the source and/or observer are moving relative to each
other. If the two are approaching, then the frequency of the wave received by the
observer will be greater than that of the emitted wave; if they move away from each
other, the wave is stretched i.e. its frequency decreases and its wavelength increases.
This effect is applicable for the light emitted by celestial objects and can be seen as a
small shift of the spectral lines in the objects spectrum. For example, if we look for some
particular spectral lines in the suns spectrum, e. g. the sodium doublet in the yelloworange portion of the spectrum, we will find that the positions of these lines and all the
other spectral lines are shifted by a small amount; caused by the following factors
(Robinson, 1981):
(a)

The rotation of Earth around its own axis that lead to a Doppler

shift of spectral lines, depending on the magnitude and direction of the velocity
10

component that is parallel to the Earth-Sun line. In the morning the observer is
on the side of Earth where the radial component of rotation is toward the sun, so
we expect to observe a shift toward the higher frequencies; a blue shift, whereas
in the afternoon the observer moves away from the sun, so the shift is toward the
red. The shift resulted from this particular motion is extremely small (at most 1.4
parts in a million).
(b)

The constant motion of the suns surface layer due to the

rotational motion of the solar rotation which has an equatorial velocity of 2000
m/s (Hathaway, 2014) cause a red or blue shift of the spectral lines, depending
on which part of the sun we are looking at.
(c)

Einsteins theory of general relativity predicts that there will be a

gravitational redshift by an amount

, where z is the gravitational

redshift, G is the Newtons gravitational constant, c is the speed of light and M


and R are the mass and radius of the sun respectively. This shift raised due to the
fact that a proton has to spend energy in order to climb out of the field of the sun,
but at the same time it must remain travelling at the speed of light, so to keep
within these two restrictions, the energy loss occurs as a change in the frequency
instead of a change in speed. Since energy is proportional to the frequency, the
frequency of the photon decreases as it escapes the sun, and this can be seen as a
shift toward the lower frequencies, or red end in the electromagnetic spectrum.
This effect gives rise to a fractional shift in frequency (

) of about one

part in a million.
(d)

Now suppose that we are talking about a star other than the sun; a

star in our own galaxy, orbiting its centre, but unlike the sun, it weaves up and
down through the galactic plane as it goes around the galaxy.
The net shift of frequency due to the above effects is of the order of 10 3.
However, for an integrated spectrum of another galaxy, one would expect to observe an
additional shift due to its relative motion with respect to the Milky Way. The shift could
be in either direction (toward the red or blue end) depending on the radial component of
the motion the galaxy, and this is exactly what has been observed for the nearby galaxies
11

so far. For example the spectrum of the Andromeda shows a slight blue shift because it
is moving toward the Milky way at about 250,000 miles per hour (111.750 km/s)
(Garner, 2012), while some other nearby galaxies show redshifts because they are
receding from us. However for more distant galaxies, this isnt often the case, as in
about the year 1912 an American astronomer called Vesto Slipher at Lowell
Observatory in Arizona noticed that almost all the frequency shifts of what were then
called spiral nebulae are redshifts, indicating that the corresponding objects appeared to
be moving away from us (Bartusiak, 2011). In 1929, Edwin Hubble combined his own
distance measurements that are based on Cepheid stars with Sliphers measurements of
the redshifts velocity and discovered that there is a linear proportionality between the
objects distance and the velocity at which it moves away from us. This relationship,
later called Hubbles law, indicates that the universe is expanding and that the redshift
phenomenon is due solely to this expansion and its measurements have now become the
most important tool in mapping the universe. Hubbles law and redshift measurements
will be discussed in more detail in the next sections.
3.2. Redshift Measurements
Usually astronomers measure the redshift of a galaxy or star by comparing its
observed spectrum with the spectrum they would expected based on its chemical
composition.

This can be done using various techniques such as photometry and

spectroscopy.
The photometric technique was developed and first described by Baum in the
sixties (Baum, 1962) and is still favoured by many Large Scale Surveys due to its
efficiency in detecting faint objects and short observing time. In photometry, the light
from a galaxy is filtered and measured across specific regions in the electromagnetic
spectrum, usually divided in a color base (e.g. ultraviolet, blue green, red, .. etc.) . Then,
the observed colors are compared to predictions from galaxy spectral energy for various
galaxy types to determine the redshift.
The spectroscopic redshift of a celestial object can be obtained by observing its
spectrum and identifying the spectral lines that correspond to specific elements and
measure the shifts of these lines with respect to their expected positions, as measured in
a laboratory on Earth. In comparison with photometry, the spectroscopic determination
12

of redshifts is more difficult and more time consuming, but it provides more reliable
data. However, sometimes a small sample of spectroscopic redshifts is used to calibrate
a large sample of photometric redshifts to reduce the error of the latter.
Astronomers usually express the wavelength (or frequency) change by the
dimensionless quantity redshift z, which is defined as (Appenzeller, 2009)

Where

is the wavelength of the detected light and

is the rest

wavelength that is determined by experiments in a laboratory or by quantum mechanics


calculations. One should notice that z can either be positive or negative depending on
the direction of the spectrum shift, or on the direction of motion of the celestial object in
the first place.

Using the relation between the frequency,

and the wavelength,

, where c is the speed of light), we can rewrite equation (3.1.a) in terms of


frequency as follows

Where

and

are the frequencies of the observed and emitted waves

respectively. In general, when dealing with objects that are in a non-relativistic motion
(

)*, the product of z and the speed of light, c, gives the radial velocity between the

object and the observer,

For very distant objects, the radial velocity (or sometimes called the redshift
velocity) is nothing but the velocity that the source is moving away from us due to the
expansion of the universe, called the recessional,

. In other words, for large distances

the recession velocity contributes the most to the redshift and any deviation from this
velocity contributes to whats called peculiar velocity.
3.3. Discovery of the Hubbles Law
In the early twentieth century, astronomers wondered about the physical nature
of the cloudy band they see in the night sky, the spiral nebulae. Many of them
*

The exact formula that takes into account the theory of special relativity and must be applied for large

distances (high redshifts) is:

, where

is the recession velocity (Cosmological Redshift).

13

believed that the spiral nebulae constituted solar systems in early stages of evolution
(O'Raifeartaigh, 2013). Others thought that they are galaxies in their own right, and are
very distant from our galaxy. This has motivated the young American astronomer, Vesto
Slipher to perform a study on the spectrum of the light from the Andromeda nebula
using a disposal 24-inch telescope at Lowell Observatory in Arizona. By 1912, Slipher
succeeded to obtain the spectra of 41 spiral galaxies (Thompson, 2011) and noticed that
almost all of them are redshifted. With the aid of Sliphers data, in 1922 Carl Wilhelm
Wirtz published a paper in which he concluded that there is a correlation between the
redshifts and the apparent brightness and the angular diameters of the observed galaxies.
Two years later, he argued correctly that since the apparent brightness and the angular
diameter of galaxies decrease with the distance, then the observed correlation must be
due to a possible redshift-distance relationship of galaxies (Appenzeller, 2009). These
were considered to be the first steps toward the discovery of the velocity-distance
relationship. However, Wirtz wasnt able to formulate the relationship because the
distance measurement methods that he was applying gave the relative distances only.
Meanwhile, Edwin Hubble was working on Mount Wilson in California and using a
100-inch telescope, called the Hooker Telescope, which was the most advanced
technology of the time to develop a reliable method to measure large cosmic distances.
He

focused

his

work

on

spiral

nebulae,

including

the Andromeda

Nebula and Triangulum (Smith, 2013), specifically concerning himself with a special
class of stars known as Cepheid variables. From their pulsating periods, Hubble was
able to obtain the absolute luminosity of the stars using the period-luminosity relation
accepted at that time and by comparing their brightness with their luminosities, he
calculated their distances. According to Hubbles calculations, the stars and the galaxies
they are a part of were much farther away than anyone had ever imagined, and the
universe was much larger than the Milky Way. In the next few years, Hubble continued
to study distant galaxies, measuring the distances of 33 galaxies for which their redshifts
were already measured by Slipher (Gribbin, 1998). He compared the redshifts and the
distances by plotting a graph of velocity (redshift) against distance, later called the
Hubble diagram (see Figure. 3.1) and noticed that the points lay on a straight line, which
indicates that velocity must be directly proportional to the distance.
14

Figure 3.1: The original Hubble diagram of 1929. The plot gives the observed redshift (again
-1
expressed as a Doppler shift given in km s ) as a function of distance in (parsec). (Appenzeller,
2009).

In 1929, Hubble published the redshift-distance correlation which is now called


the Hubbles Law and summarised by the following formula (Robinson, 1981)

Where the constant of proportionality,


and

is called the Hubble constant and

are the radial velocity and the proper distance* of the receding galaxy respectively.

This relationship implies that a galaxy twice as far from us than another galaxy is seen to
be receding twice as fast as the nearer galaxy. This conclusion can be considered as the
foundation stone of the expanding universe model and the Big Bang theory.
3.4. Hubbles Constant
In fact, the constant of proportionality in Hubbles Law is not really constant
over time because it does depend upon time, therefore, it should probably be called
Hubble parameter rather than Hubble constant. However, its usually written with a
subscript 0 to denote that it is the value of Hubble parameter at present time.
It is of high importance in cosmology to get an accurate value for the Hubble
constant because it affects the accuracy of determining the size and age of the universe,

The proper distance is a distance between two nearby events in the frame in which they happen at the
same time (Michel-Marie Deza, 2006) and can change over time, unlike the comoving distance which is
constant over time.

15

but it is somewhat challenging. To obtain the value of the Hubble constant, one needs to
have sets of measurements. The first is the galaxies redshifts (which yield their radial
velocities). The second measurement is one of the most difficult measurements to make
in astronomy; the distance measurements. In addition to this, the sample of galaxies that
are used for the determination of H must be comprised of objects that are far enough
away to avoid the contribution of the peculiar velocities (local motions) on the net radial
velocity.

Due to these difficulties, the first values of the Hubble constant were

completely inaccurate; for example, the numerical value of the slope of the original
Hubbles diagram was about seven times too large due to a systematic error he had in his
distance data (Appenzeller, 2009). In the next 30 years, the Hubble constant reduced to
about half of its current value, then the relative error has been reduced to about 10%
during the rest of the 20th century (Li, Qin, & Zhang, 2014) (Trimble, 1996). The most
recent value for the Hubble constant is (67.800.77) (km/s)/Mpc, which is obtained by
the Planck Mission and published in March 21, 2013 (Ade , et al., 2013). The Planck
Mission, lunched on May 14, 2009, is a project operated by the European Space Agency
(ESA) and aims to scan the Cosmic Microwave Background radiaton (CMB) using the
dedicated European Space Agencys Planck satellite. As the microwave radiation is
known to be the oldest light in the universe, Plancks measurements are expected to
provide a wealth of information about the early universe and its subsequent evolution.
Specifically, the data can be used to set tight constraints on cosmological parameters and
the ionization history of the universe. Moreover, it can be used to probe the dynamics of
the inhomogeneities produced in the inflationary era and to test the fundamental physics
beyond the inflation. Figure 3.2 illustrates the map of the cosmic microwave background
radiation yielded by the Planck mission.

16

17

Figure 3.2: Cosmic Microwave Background seen by Planck. From (Cosmic Microwave Background seen by Planck, 2013).

Chapter 4: The Redshift Surveys, Past, Present and Future


The first three-dimensional maps of the universe that combined the
distance information, obtained from the redshifts, with the angular information
obtained from locating galaxies on the celestial sphere were those of Gregory &
Thompson (Stephen A. Gregory & Laird A. Thomson, 1978) and Gregory, Thompson,
& Tifft (Gregory, Thompspn, & Tifft, 1981). They were pencil beam surveys targeting
specific clusters such as Coma, A1367 and Perseus clusters, with the purpose of
identifying superclusters.
Today, there are numerous redshift surveys, the number of which is growing, for
two reasons.

The first is that technological advancement in the detectors and

spectrographs employed for the gathering of redshift data. The second is that redshift
surveys have proven that they play a key role in studying the distribution of matter in the
universe and the evolution of the large-scale structure.
This chapter consists of eleven sections; the first is a brief introduction to the
methods of redshift surveys, the following ten sections will highlight the redshift surveys
that are used in this project.
4.1. The Survey Methods
In most cases, the samples of the redshift surveys are selected using quantifiable
selection criteria; i. e the sample is defined as limited by some photometric property,
usually received flux or by limiting angular diameter.
This section introduces the different classes of redshift surveys, classified
according to the selection procedure they follow.
4.1.1. Pencil-beam Surveys. The first redshift surveys that were performed to
study the large-scale structure of the universe i. e. Gregory & Thompson and Gregory
(1978), Thompson, & Tifft (1981) were pencil-beam surveys, where a small area of the
sky that is delimited by specific ranges of the angular coordinates is selected and
intensive redshift measurements are made on it. The pencil-beam method is considered
the most economical method since it can yield information on a great depth and achieve
completeness and limited telescope time simultaneously. The 2dF survey (shown in
Figure. 4.2) is an example of a pencil-beam survey.

18

Figure 4.1: The first set of observations done for the CFA redshift survey in 1985 by Valerie de Lapparent,
Margaret Geller and John Huchra (Huchra J. ).

4.1.2. Slice Surveys. In this method, galaxies are observed in a thin strip of
the sky with a specified range of declination (from 1 as in CFA1 (Figure. 4.1) to 6 as
first wide slice of a deeper extension of the original CfA survey). The radial coordinate
is the redshift and right ascension is the angular coordinate. The Slice surveys offer
the required resolution and depth for testing the geometry of the galaxian distribution.
4.1.3. All-Sky and Large Surveys. In contrast to the methods mentioned above,
the large surveys extend to large values along all of the three dimensions (RA, DEC and
redshift) to cover a huge volume of the sky. The selection criteria in the large surveys
may be based on the galactic extinction, sky coverage of the telescope used, and limits
of the target catalogue, usually in apparent magnitude, flux, or diameter, and may also
be restricted by morphology or surface brightness constraints on detectability. The
efficiency of each survey is then affected by the biases introduced by selection
procedures (Giovanelli & Haynes, 1988).
4.1.4. Targeted Surveys with special kinds of objects. In this method, targeted
observations are carried out to learn about the properties of galaxies, and hence provide
an important feedback that improves our understanding of the formation and evolution
of the large-scale structure. The surveys may be targeted toward clusters of galaxies,
19

active and unusual galaxies of various kinds, or pairs and groups of galaxies. However,
this kind of survey does not provide the sufficient information for studying the largescale structure features or layout (Strauss & Willick, 1995).
4.1.5. Blind Redshift Surveys carried out in the 21 cm line of natural
hydrogen (HI). These surveys study the redshift of the 21 cm emission line, which
corresponds to the hyperfine transition of electron in the hydrogen atom between the
parallel spin (triplet) state which has higher energy to the anti-parallel spins (singlet)
state which has lower energy. The strength of the signal at the emission or absorption
lines depends on the relative occupation number of the ground state and the occupation
number of the excited state; hence it gives information about the density of the neutral
hydrogen along the line of sight. This type of survey is one of the fundamental tools for
studying the formation of the first structures during the dark ages and the Epoch of
Reionization (Rhee, et al., 2013).
However, none of the surveys that are used in this project are of the last two
kinds of the above mentioned surveys.
4.2. Sloan Digital Sky Survey (SDSS)
The Sloan Digital Sky Survey is a significant project to produce detailed, 3dimensional maps of more than a quarter of the entire sky, using a dedicated 2.5-meter
instrument at the Apache Point Observatory in New Mexico (Lahav & Suto, 2004). The
Survey began in 2000 and started to release data to the scientific community and the
general public periodically.
In its first two operational stages (SDSS-I, 2000-2005, SDSS-II, 2005-2008), the
project obtained comprehensive multi-colour images of more than a quarter of the sky
that created 3-dimensional maps containing more than 930,000 galaxies and more than
120,000 quasars. The Sloan Digital Sky Survey is continuing through the Third Sloan
Digital Sky Survey (SDSS-III) that started observations in mid-2008 and will continue
until 2014. This phase consists of four distinct surveys, carried out using the same
facilities.

The four surveys are: APO Galactic Evolution Experiment (APOGEE),

Baryon Oscillation Spectroscopic Survey (BOSS), Multi-object APO Radial Velocity


Exoplanet Large-area Survey (MARVELS) and the Sloan Extension for Galactic

20

Understanding and Exploration (SEGUE-2). Each of the SDSS-III surveys is designed


for different purposes and operates differently.
The first data collection of SDSS-III (Data Release 8) was released in January
2008. It contains all of the imaging data taken by the SDSS imaging camera which
covers about 14,000 square degrees of the sky in addition to the new spectra taken by the
SDSS spectrograph during its last year of operation for the SEGUE-2 project and
includes imaging of roughly 5200 square degrees.
The second SDSS-III data collection (Data Release 9) was released in August
2012. It contains the first release of BOSS spectroscopy to the public and also includes
all imaging and spectra from previous data releases with some important updates such as
providing corrected astrometry for the imaging Data Release 8 by making some
improvements to the astrometric calibration. Data Release 9 covers over 3,300 square
degrees. The median redshift of the SDSS is z=0.1 and the maximum detected redshift
for galaxies is z=0.7 and foe quasars is z=5; and the imaging survey is supposed to
include quasars as far as z=6.
The newest release is Data Release 10 which was completed in July 2014 and
contains the first release APOGEE infrared Galactic spectroscopy as well as hundreds of
thousands of new galaxy and quasar spectra from the BOSS survey and some
modifications to the Data Release 9. The total sky coverage at completion of Data
Release 10 is 14,555 square degrees and optical spectra measurements for 1,848,851
galaxies, 308,377 quasars and 736,484 stellar objects as well as 57,454 infrared stellar
spectra. The SDSS-III will continue operating and releasing data and hopefully be
complete in few more years. The obtained SDSS image is considered to be the largest
colour image of the sky ever made.
The project was managed by the Astrophysical Consortium for Participating
Institutions and named after Alfred P. Sloan Foundation who provided the major Fund
(Sloan Digital Sky Surveys, 2014) (York, et al., 2000).
4.3. Two-degree-Field Galaxy Redshift Survey (2dFGRS)
The Two-degree-Field Galaxy Redshift Survey (2dFGRS) is a major
spectroscopic redshift survey conducted on the 3.4-meter Anglo-Australian Telescope in
the Siding Spring Observatory in Australia. The project started in 1997 and by 11 April
21

2002, the project was completed with spectra measurements of 245591 objects brighter
than magnitude of bJ=19.45, 221414 of them yielded reliable redshifts of galaxies, which
makes it to be considered as the second largest redshift survey next to the Sloan Digital
Sky Survey. The redshifts of almost all galaxies is less than z=0.3 with a median of z =
0.11. The total sky area covered by 2dF survey is approximately 1500 square degrees,
distributed in three regions: two declination strips, in the northern and southern galactic
hemispheres, and 100 random fields scattered around the southern galactic hemisphere
strip as shown in figure 4.2 (Colless, et al., 2003).

Figure 4.2: The distribution of 63000 2dFGRS galaxies in the NGP (left panel) and SGP (right panel) strips
(Lahav & Suto, 2004).

4.4. Six-degree-Field Galaxy Redshift Survey (6dFGRS)


The Six-degree-Field Galaxy Redshift Survey (6dFGRS) is aimed to measure
redshifts and peculiar velocities over almost the entire southern sky. The survey was
conducted on the Anglo-Australian Observatorys 1.2-meter UK Schmidt Telescope
between 2001 and 2009. The survey achieved 136,304 spectra measurements that
yielded 110,256 reliable and unique redshifts with median of z=0.053. The survey is the
largest redshift survey of the nearby universe (z 0.15) and the third largest survey next
22

to 2dFGRS and SDSS. The 6dFGRS covers an area of 17000 square degrees of the
southern sky with |b| < 10 which is more than ten times the sky area of the 2dFGRS and
almost twice that of the SDSS (Jones, et al., 2009).
4.5. VIMOS-VLT Deep Survey (VVDS)
The VIMOS-VLT Deep Survey (VVDS) is a comprehensive imaging and
spectroscopic redshift survey, carried out using the VIMOS spectrograph which is built
by the VIRMOS consortium of French and Italian. The aim of the project is to provide a
complete picture of galaxy formation and evolution of large scale structure growth of the
universe over a very broad redshift range (0 < z 6.7) over sixteen square degrees of the
sky in four separate fields. The VVDS is a combination of three nested surveys, selected
on the basis of apparent magnitude as follows: Wide (17.5 iAB 22.5; 8.6 square
degrees), Deep (17.5 iAB 24; 0.6 square degrees) and Ultra-Deep (23 iAB 24.75;
512 square arcmin) with a total of 22434, 12051, 1041 measured galaxy redshifts
respectively. In the redshift scale, the VVDS succeed in measuring high redshifts with
4669 redshifts in 1 z 2, 561 redshifts in 2 z 3 and 468 with z >3 (Le Fvre, et al.,
2005).
4.6. The Deep Extragalactic Evolutionary Probe project (DEEP)
The Deep Survey is a large scale survey of distant galaxies conducted to study
the formation and evolution of galaxies, the origin of large-scale structure, the nature of
the dark matter, and the geometry of the Universe using the twin 10-m W.M. Keck
Telescopes and the Hubble Space Telescope (HST).
DEEP project is divided into two phases, the first (DEEP I) used the Low
Resolution Imaging Spectrograph (LRIS) to study a sample of ~ 1000 galaxies to a limit
of I = 24.5 and a median redshift of about unity. Phase 1 was designed to examine the
technical feasibility and establish the scientific scope of the program, prior to conducting
phase 2. The second phase (DEEP II) started in spring 2002 and used the largest
spectrographic detector of its type ever made, which is the Deep Multi-Object
Spectrograph (DEIMOS) that is capable of observing 140 galaxies at a time. The goal
of DEEP II is to include a rage sample of greater than 50000 galaxies brighter than Iband magnitude 23.5 with high resolution and in the pre-selected redshift range of z =
0.7-1.55 (Davis, et al., 2003).
23

4.7. Two Micron All-Sky Survey (MASS)


The Two Micron All-Sky Survey (MASS) survey was conducted between 1997
and 2001 and covered almost the whole sky (99.998% of the celestial sphere) in three
near-infrared band. Observations were conducted from two dedicated infrared 1.3-meter
m diameter telescopes located at Mount Hopkins, Arizona, and Cerro Tololo, Chile. The
observed wavelengths were 1.25, 1,65 and 2.16 microns.
The project was a collaboration between The University of Massachusetts that
constructed and maintained the observatory facilities and operated the survey, and the
Infrared Processing and Analysis Center (JPL/ Caltech) that had the responsibility of the
data processing and data product generation.
The 2MASS data were release for general public in 2003 release and included
4.1 million compressed FITS images covering the entire sky, 471 million source
extractions in a Point Source Catalogue, and 1.6 million objects identified as extended in
an Extended Source catalogue (Huchra, et al., 2012).
4.8. VIMOS Public Extragalactic Redshift Survey (VIPERS)
VIMOS Public Extragalactic Redshift Survey (VIPERS) is an on-going survey
started in 2008 and carried out using the European Southern Observatorys Very Large
Telescope (ESO VLT). The aim of the project is to build a large sample of ~100,000
galaxies with red magnitude I(AB) brighter than 22.5 over an area of 24 square degrees.
The survey is focused on a long redshift range (0.5 < z < 1.2), hence yielding a large
volume (5 x 107 h-3 Mpc3). The first VIPERS data release includes spectroscopic
measurements for 57204 objects, 54756 of them are galaxies (Franzetti).
4.9. Galaxy And Mass Assembly Survey (GAMA)
The Galaxy And Mass Assembly Survey (GAMA) is a project to study the
structure on scales of 1 Kpc to 1 Mpc. This wide range will be very helpful in studying
the galaxy evolution and the large-scale structure.
GAMA surveys will be carried out using a number of instruments: The AngloAustralian Telescope (AAT), the VLT Survey Telescope (VST), the Visible and Infrared
Survey Telescope for Astronomy (VISTA), the Australian Square Kilometre Array
Pathfinder (ASKAP), the Herschel Space Observatory and the Galaxy Evolution
Explorer (GALEX).
24

The newest GAMA data release (DR2) gives spectra, redshifts and abundance of
information for 72,225 objects from the first phase of the GAMA survey (2008 - 2010,
usually referred to as GAMA I) (Baldry, et al., 2010).
4.10. FORS Deep Field (FDF) spectroscopic survey
Using the visual and near UV FOcal Reducer and low dispersion Spectrograph
(FORS) instrument, the FORS Deep Field (FDF) spectroscopic survey was performed to
improve our understanding of the formation and evolution of galaxies in the young
Universe. The survey began in 1999 and covered a small angular area, 7 7, just
about the size of a large galaxy cluster at high redshift and the targets were selected in
the basis of photometric redshifts. The survey detected 341 reliable redshifts, 98 of them
correspond to starburst galaxies and QSOs at z > 2. The redshift of the observed
extragalactic objects ranges between 0.1 and 5.0 (Noll, et al., 2004).
4.11. Team Keck Redshift Survey (TKRS)
Team Keck Redshift Survey (TKRS) is a project conducted by W.M. Keck
Observatory and used the Keck II telescope to observe the spectra of nearly 3000
sources, yielding secure spectroscopic redshifts for 1536 objects (1437 are galaxies).
The redshifts are accurate to 100 km/sec and the limiting magnitude is R=24.3. The
TKRS results enabled numerous studies of large-scale structure, galaxy dynamics, and
abundances (Wirth, et al., 2004).

25

Chapter 5: Data Collection and Statistical analysis


As the aim of this project is to study the distribution of galaxies throughout
universe without focusing on a specific region of the sky or focusing upon a certain type
of celestial object, it was decided that the publicly available data of various Redshift Sky
Surveys would be used that utilise different selection criteria. However, this required
the following of certain procedures to check for data reliability and consistency and for
the existence of duplications, if there should be any. The data was collected from the
following ten Redshift Sky Surveys: Sloan Digital Sky Survey (SDSS, 10th Release),
Two Micron All-Sky Survey (2MASS, the final data product), Two-degree-Field Galaxy
Redshift Survey (2dFGRS, the best spectroscopic observations of the final release), Sixdegree-Field Galaxy Redshift Survey (6dFGRS, Data Release 3), Galaxy And Mass
Assembly survey (GAMA, Data Release 2), VIMOS Public Extragalactic Redshift
Survey (VIPERS, PDR-1), The Deep Extragalactic Evolutionary Probe project (DEEP,
Data Release 4), VIMOS-VLT Deep Survey (VVDS, First Epoch sample), Team Keck
Redshift Survey (TKRS) and FORS Deep Field spectroscopic Survey (FDF).
This chapter describes the procedures implemented to check that the data
collected from the various databases is consistent and that the two tests performed to
check whether the distribution of the radial dimension, z is periodic.
5.1 Checking data consistency
The high quality and reliable redshifts are extracted from each survey data
collection, such that

< 10%, or measurement confidence 90%.

In all of the survey catalogues used in this project, the positions of the celestial
objects are given with respect to the J2000 equatorial coordinates. However, we wanted
to examine the periodicity hypothesis in the distribution of the co-moving radial
distance, as well as in the distribution of the z in the equatorial coordinates there arose a
need to perform a conversion from the J2000 equatorial coordinates to the co-moving
coordinates. Moreover, the angular dimensions were converted from the equatorial
coordinates to the galactocentric coordinates so they can be used in any further studies.
The next two subsections discuss the procedures implemented to generate reliable data
sets in different coordinate systems from the original source catalogues.

26

5.1.1 Coordinate Transformations. The equatorial coordinate system originates


at the center of Earth and its reference plane is the Earths equatorial plane which is the
projection of the earths equator on the celestial sphere.

Based on the equatorial

coordinates, the angular position of any celestial object can be described by two
quantities: right ascension; RA, and declination, Dec. The right ascension is measured
eastward along the celestial equator starting from the vernal equinox. It could be given
in hour angle and its subdivisions (from 0h0m 0s which corresponds to the direction of
the vernal equinox, to 24

0m 0s) or in degrees (from 0 to 360 ). The declination

measures the angular distance of an object perpendicular to the celestial equator, and
extends from 0 at the celestial equator to +90 at the north celestial pole and down to 90 at the south celestial pole. For the purposes of consistency, the angular dimensions
(RA and Dec) of all of the objects in the data collection have been converted to decimal
degrees.
It must be kept in mind that the orientation of the equator and the ecliptic are not
fixed, but rather they are changing with time due to the precisional motion of the earth.
Thus, a standard reference frame is usually based on the mean equator, the mean
ecliptic, and the equinox of some dedicated epoch. The most commonly used is the
J2000.0 equatorial coordinates which is based on the Julian date calendar and the right
ascension and the declination are defined from the mean vernal equinox and the mean
equatorial plane at 12:00 UTC on January 1st in the year 2000 (Takahashi, Kondo,
Takahashi, & Koyama, 2000).
The radial dimension of an object in the equatorial coordinates is the component
along the line of sight between the observer and the position of the object, measured at a
specific moment. While this distance (called the proper distance) is changing with time
due to the expansion of the universe, there is another coordinate choice, called the
comoving coordinates, which allow us to express the radial distance to an object with
eliminating the increase in the distance due to the expansion. In other words, in the
comoving coordinates the distance to an object (called the comoving distance) is
constant over time despite the expansion of the universe.

However, the detailed

definition of the comoving coordinates is beyond the scope of this project, and what we
should concern ourselves with is the conversion of the redshift, z, to the comoving
27

Figure 5.1: The commoving distance as a function of z, based on equation 5.1.

distance. This conversion task has been accomplished with the aid of a MATLAB
function that is available online by Instituto de Fsica y Matemtica (Surendran, 2011)
The code is based on the following formula

Where

the comoving radial distance, z is the redshift value that we want to convert,

is the speed of light and

is the Hubble constant. The numbers 0.27 and 0.73 refer

to the fractions of the matter and dark energy respectively. Using the above formula, a
graph or r versus z has been plotted to be used in the reverse conversion; from comoving
distance to redshift (the graph is shown in Figure. 5.1). This relation tells that a fixed
separation in the z scale corresponds to a decreasing spacing in comoving distance scale;
e. g. a spacing of 0.1 between z=1 and z=1.1 is equivalent to ~242 Mpc in the comoving
distance scale, while the same separation in the z scale between z=4 and z=4.1 is
equivalent to ~70 Mpc in the comoving coordinates.
Furthermore, the measurements have been converted from the equatorial
coordinates to the galactic coordinate system. The galactic coordinates are based on the
galactic plane and centered on the sun. The latitude angle of the galactic coordinates is
denoted by the letter b, and its measured from 0 to 90 northward and from 0 to -90
28

southward the galactic equator. The galactic longitude is represented by the letter , and
is measured in the galactic plane using an axis pointing from the sun to the galactic
center (where = 0 ) and extends counter clockwise as seen from the north galactic pole
to 360

The conversion from the equatorial coordinate system to the galactic coordinate

system has been accomplished with the aid of a Fortran 95 code emailed to us by
Professor Matthew Colless, which uses the following set of equations:
[

Where Ra, Dec, b and are all given in degrees.


Moreover, in some researches such as (Napier & Guthrie, 1997) the existence of
a redshift periodicity has been detected in the galactocentric frame of reference, which
has the galactic center as its origin. However, we have performed a simple test to
determine if whether the difference between the redshifts in the galactocentric
coordinates; zgalactocentric, and the redshifts in the equatorial coordinates (zequatorial) is
significant. The test was as follows: If the galactic center, Earth and the targeted
celestial object are approximated to form a right tringle, then the difference between
zgalactocentric and zequatorial can be written as,
|

|
Where

is the galactic center distance from Earth. Using

~ 8.33 kpc

(Gillessen, et al., 2009), the difference can be calculated for a range of values of
as illustrated in Figure 5.2. From the calculations we have concluded that the
difference is too small and insignificant.
5.1.2 Duplication checking
Since we are dealing with a large data set (for ~ 3 million objects), comparing the
coordinates of each object with the coordinates of all other objects would require doing
very heavy computations. To reduce the number of calculations, we have made use of
the following method: the data lines have been ordered and then sorted into 36 groups,
29

based on the right ascension values. The three position coordinates (RA, Dec and z) of
each object have then been compared to all of the other objects of the same group that
the object belongs to. Two objects are considered to be duplicate if their redshifts

Figure 5.2: The difference between the redshifts with respect to the equatorial and galactic coordinates.

and

) match precisely and the difference between their angular dimensions doesn't

exceed the approximated angular diameter. These conditions condition can be written as

(
Where

is the galactic diameter and is approximated to be fixed as the diameter

of the Andromeda which is 110 kly (Kambic, 2009) . However, this checking method is
valid for low redshifts only (z << 1).
5.2 Checking for periodicity
The primary statistical method employed in this study for the detection of
periodicity is the periodogram. Before calculating the periodogram, the distribution of
the redshifts was represented in a simple frequency histogram in order to get a rough
indication of periodicity, and also to determine the most populated range of redshifts in
each survey catalogue. The two methods are described in the following subsections.

30

5.2.1 Histogram
For Large sets of data, the histogram is an efficient method to use in order to
derive various features of the distribution. It uses rectangles to represent the counts of
the data point falling in various ranges. The most common form of the histogram can be
obtained by dividing the data range into equal-sized bins. Then the number of data
points that fall into each bin is counted and plotted against the corresponding interval.
The shape of the histogram and the data features that can be extracted from it are
highly dependent on the bin size. If the bins are wide, the noise due to the sampling
randomness can be reduced, but at the same time, important information might be
omitted. On the other hand, using narrow bins may cause random variation effects to
appear as meaningful information, but it gives greater precision to the density estimation
as a benefit. Therefore, its advisable to try different bin sizes in order to obtain a more
complete picture of the data. In this study, the histogram of the redshift distribution has
been computed for four different bin sizes: 0.2, 0.1, 0.01 and 0.001.
5.2.2 Periodogram
A Periodogram is an efficient statistical tool for examining the periodic (cyclic)
behavior in a time series, where it identifies the dominant periods (or frequencies) of the
sequence. Assume we have a time series of length N; { }

,a

smoothed periodogram of the series can be obtained by carrying out the following main
three steps:
1. Multiplying the sequence by a window function to reduce the noise and the
effect of the leakage in the Fourier transform due to sudden changes in the
start and end of the data (Kang, 2008).
2. Computing the discrete Fourier transform (DFT) which is defined as (Vio,
Diaz-Trigo, & Andreani, 2013):

3. Averaging the squared absolute value of the DFT of the signal to obtain the
periodogram,

31

| |
A plot of

s as a function of frequency is also called the periodogram (also the

power spectrum) of the series, and the analysis is called the periodogram analysis
(Chatfield, 2004). If there is a strong sinusoidal signal in the time series with some
frequency, then there will be a peak in the periodogram at that frequency. If the periodic
signal is non-sinusoidal, then there will be a fundamental peak in the periodogram at the
frequency of the signal but also at multiples of that frequency.
However, if the sampling rate,

(which is the reverse of the sampling step) is

chosen to be very small, then a phenomenon called aliasing may occur.


Aliasing refers to when the frequencies above a certain frequency, called the
) will, after sampling, be indistinguishable from

Nyquist frequency (

frequencies below the Nyquist frequency and distort the spectrum. In order to void the
aliasing problem, one can filter out components whose frequency exceeds the Nyquist
frequency prior to sampling. Another solution worth mentioning is to choose a sampling
rate that is sufficiently high (much higher than the frequency of interest) so that the
frequency components produced by aliasing will be far away from the range of
frequency of interest and can be easily discriminated. More details about the aliasing
phenomenon and its solutions can be found in (Hearn & Metcalfe, 1995).
For testing the white noise, one can use the cumulative periodogram which
defined as (Diggle & Fisher, 1991)

For a purely random series, the plot of

s against

would consist of

a straight line joining (0, 0) and (0.5, 1). If a series has a periodicity with some
frequency, then there will be a jump in the cumulative periodogram at that frequency
(where the periodogram has a peak).
The code that is used in this project to calculate the periodogram, both the
classical and the cumulative periodogram, is attached in Appendix. A.

32

Chapter 6: Results and discussion


A total of 2841496 redshift measurements were obtained from ten redshift survey
catalogues. After filtering the redshifts to extract the most reliable values, the number is
reduced to 2763064 redshifts, combined with the angular distances of the associated
celestial objects. The collective sample covers the whole sky area (RA from 0 to 360
and Dec from -90 to 90). Table 6.1 summarizes the number of objects included in each
survey sample, as well as in the collective set before and after the filtering process.
Table 6.1: Summery of the redshift measurements before and after filtering process
Sample

Number of redshifts (before filtering) Number of reliable redshifts

SDSS

1982640

1979282

2MASS

240510

240291

2dF

245591

227190

6dF

124647

124326

GAMA

90409

83610

VIPERS

61221

45156

DEEP

50319

35752

VVDS

40081

25685

TKRS

5737

1473

FDF

341

299

2841496

2763064

Collective sample

Note that the number of reliable redshifts in the collective sample is not the sum
of the numbers of reliable redshifts of the individual sample, because the collective
sample has been subjected to the same filtering process in order to detect the duplicate
objects between the different survey samples.
In this chapter, we present a discussion of the radial distribution of the celestial
objects using the histograms and the periodograms that we have obtained. We also
discuss the main conclusions of the study.
Note that all of the plots have been obtained using MATLAB 2014b, while the
calculations (Filtering, conversions, statistics, .. etc.) have been done using Fortran 95 as
well as MATLAB 2014b.
33

6.1 Radial Distributions


The distributions of the celestial objects along the radial dimension, z, have been
examined using histograms and periodograms for different bin sizes and different
sampling rates. The next two sections will present the resulted diagrams and the general
conclusions that can be drawn from them.
6.1.1 Histograms.
The z values of the collective sample have been presented using histograms with four
different bin widths: 0.2, 0.1, 0.01 and 0.001, as shown in figure 6.1. Comparing the
four histograms, one can see clearly that the 0.2 and 0.1 bin sizes are too large such that
the resolution of the resultant histograms are not good enough to extract the features of
the distribution. In other words, most of the data fall in the first few bars so we may lose
some details of the shape of the distribution. Also, the 0.001 bin size is too small
because it shows unimportant fluctuations that could have arisen from the error of the
measurements (since 0.001 is comparable with the error of the small values of z, see
section 5.1). The comparison between the histograms with the four different bin sizes is
represented in Figure 6.1.
Most of the z values fall in the interval (0 z 4) as can be seen from figure 6.2.
This can be further clarified by representing the frequency (or number of counts) in the
logarithmic scale as shown in Figure 6.3. The lack of measurements in the higher region
(z 4), which is due to the limitations of the observational techniques, lead us to rely
more on the (0 z 4) region in drawing our conclusions than on the higher region.
The maximum z value in the collective sample is 7.054 which is obtained from
the SDSS survey.
Among the ten survey samples, the SDSS sample contributes the highest in the
collective sample (~ 72%). Figures 6.4 and 6.5 represent the histogram of the SDSS
data with bin size of 0.01 using a linear scale and logarithmic scale for the vertical axis
respectively. The histograms show that the well-sampled region extends to a quit high z;
such that the frequency line is mostly above 100 even near z=4.6. However, the sample
extends to a maximum of z = 7.054.
Compared to the SDSS, the 2MASS sample is focused on the low z region,
where the maximum z is 1.968 and the frequency is less than 10 for the values above
34

0.3, while it is above 103 in the region (z 0.2), exceeding 3 104 measurements near
z=0.1.

Figures 6.6 and 6.7 show the frequency histogram of the 2MASS sample

represented using the linear scale and the logarithmic scale respectively, both with bin
size of 0.01.
Figures 6.8 and 6.9 include the histogram of the 2dF sample. This histogram
shows that the number of objects is large for low values of z, reaching ~17000
measurements near z=0.1 specifically, but it starts to drop rapidly and ends with less
than 10 measurements for each z interval in the region (z 0.4). The z values of the 2dF
data spread to z=3.500.
The 6dF data is represented in Figures 6.10 and 6.11. These figures show that
most of the measurements are in the (z 0.2) region, where the frequencies in this
region are of order 102 to 104 with a peak of about 18000 at z ~ 0.05. The maximum
value of z in this survey is 3.793.
The z values in the GAMA survey extend to 3.435; however, the numbers of
objects in the high z region (z 0.3) are negligible compared to those in the lower
region, as can be seen in Figure 6.12. The figure also shows that there is a peak in the
region around z = 0.15.
The histogram of the VIPERS data has a broad shape and the well-sampled
region (z 1.6) is nearly symmetric about z = 0.7 (see Figure 6.13). The z values of this
survey spread to z = 4.357. The VVDS sample covers the region (0 < z 4.980) and has
the highest concentration around z=0.6 as seen from Figure 6.15.

However, the

frequency bars are below 10 for the z values that are larger than 1.6.
The two samples with the least number of objects are the TKRS with 1473
objects and the FDF with 299 objects. The histograms of Figures 6.16 and 6.17 show
the redshift distribution of the galaxies in the TKRS sample and the FDF sample
respectively. From the histograms, it can be seen that in the TKRS sample the z values
are distributed randomly over the lower region and started to become very rare at z=1.53
and above. The sample covers the z region up to z=3.530. The frequency of the FDF
histogram ranges between 0 and 10 for the whole (included) range (0 < z 5.000) except
at z ~ 0.14 where it exceeds 20.

35

36

Figure 6.1: The collective sample redshifts represented using histograms with four different bin sizes: 0.2, 0.1, 0.01 and 0.001 (From up to bottom).

37

Figure 6.2: Histogram of the collective sample redshifts with bin size of 0.01.

38

Figure 6.3: Histogram of the collective sample redshifts with bin size of 0.01. The Frequency is represented using the logarithmic scale.

39

Figure 6.4: Histogram of the SDSS sample redshifts with bin size of 0.01.

40

Figure 6.5: Histogram of the SDSS sample redshifts with bin size of 0.01. The Frequency is represented using the logarithmic
scale.

41

Figure 6.6: Histogram of the 2MASS sample redshifts with bin size of 0.01.

42

Figure 6.7: Histogram of the 2MASS sample redshifts with bin size of 0.01. The Frequency is represented using the logarithmic scale.

43

Figure 6.8: Histogram of the 2dF sample redshifts with bin size of 0.01.

44

Figure 6.9: Histogram of the 2dF sample redshifts with bin size of 0.01. The Frequency is represented using the logarithmic scale.

45

Figure 6.10: Histogram of the 6dF sample redshifts with bin size of 0.01.

46

Figure 6.11: Histogram of the 6dF sample redshifts with bin size of 0.01. The Frequency is represented using the logarithmic scale.

47

Figure 6.12: Histogram of the GAMA sample redshifts with bin size of 0.01.

48

Figure 6.13: Histogram of the VIPERS sample redshifts with bin size of 0.01.

49

Figure 6.14: Histogram of the DEEP sample redshifts with bin size of 0.01.

50

Figure 6.15: Histogram of the VVDS sample redshifts with bin size of 0.01.

51

Figure 6.16: Histogram of the TKRS sample redshifts with bin size of 0.01.

52

Figure 6.17: Histogram of the FDF sample redshifts with bin size of 0.01.

6.1.2 Periodograms
In this study, the periodicity has been tested using the following three samples:
1. The collective sample (the collection of the ten survey samples) for the whole
range of z
2. The well-sampled region of the collective sample (z 4).
3. The collective sample data converted to the comoving coordinates
For each sample, the histogram of the z values has been used as the time series
and its spectrum estimate has been calculated using different sampling rates. Figure
6.18 shows the periodograms of sample 1, using histograms with different bin sizes: 0.2,
0.1, 0.01 and 0.001 respectively. All of the periodograms show peaks in the low
frequencies range at similar positions, such that there are peaks at 0.47 in the two upper
panels and at 0.49 for the lower two. The peak at 1.09 in the first upper two is close to
the 1.07 and 1.1 peaks that exist in the third and fourth periodograms respectively. Also,
the 1.6 and 2.3 peaks are present in all of the periodograms.

The three lower

periodograms have common peaks near 2.9, 3.5, 4.2 and 4.7. Common peaks continue
to exist for the rest of the frequency range; however their heights are much lower than
the heights of the first few.
These results can be further clarified using the cumulative periodograms. Figure
6.20 shows the corresponding cumulative periodograms of Figure 6.19 periodograms.
This supports the existence of the peaks at ~0.49 (or 0.47), ~1.09, 1.6 and 2.3 where
there are noticeable jumps at these values.

The peaks at higher frequencies are

difficult to be distinguished using the cumulative periodogram; since it grows to be


equal to one at the Nyquist frequency.
Figures 6.21 and 6.22 present the periodograms of the well sampled region (z
4) and their corresponding cumulative periodograms respectively. The Figures show
similarity with the pervious Figures (Figures of sample 1), in which all of the peaks
occur at positions close to those of the full sample. The 0.47 peak is present in the upper
two periodograms and there is a peak at 0.49 in the lowermost periodogram. The 1.09
peak can be clearly seen in the periodograms with sampling rates of 5 and 10, which are
close to the 1.0 peak in the other two periodograms. A peak at 1.7 and a peak close to
2.2 appear in all of the periodograms.
53

The cumulative periodograms of sample 2 support the periodicities that are


mentioned in the above paragraph, where there is some non-linearity that can be noticed
around the peak positions.
However, if we look at the peak positions in the z domain, we can easily notice
that they reflect the shape of the sample that doesnt represent the real distribution of
galaxies in space. The peak at 0.49 in the frequency domain corresponds to the peak
that is slightly above z=2 in the histogram (see Figure 6.2). Also, the peaks at 1.09, 1.6
and 2.3 in the frequency scale are equivalent to z=0.92, 0.63 and 0.43 and we can
observe 3 peaks in the histogram that their peak-base widths are of the same order of
these values.
Due to this under-sampling effect, we are unable to detect low-frequency
periodicities. However, since we are using a small bin size ( z=0.01) we can still look
for short periods (high frequency signals) that are not affected by the sample selection.
Figure 6.22 shows the periodograms of sample 1 and sample 2 represented in the
logarithmic scale. The figure helps to identify the range of frequencies that we are able
to probe. Both periodograms show that there is a turning point near f=10, where all of
the peaks below it could be due to the large bell-shaped curves of the sample. Figures
6.23 and 6.24 show the periodograms of sample 1 and sample 2 respectively, focusing
on the high-frequency range (f >10). From the figures, we can observe that there is no
special peaks that can ascribe to a periodicity in the redshift distribution.
The periodogram and the cumulative periodogram of the radial distances
expressed in the comoving coordinates are illustrated in Figures 6.25 and 6.26. Because
of the above-mentioned under-sampling effect, the Figures show fake periodicities at
low frequencies, especially at 4.9 10-4 /Mpc (equivalent to 2.0 103 Mpc in the
comoving distance scale; or, redshift of 0.52), 7.9 10-4 /Mpc (1.3 103 Mpc; or,
redshift of 0.31), 11 10-4 /Mpc (0.91 103 Mpc; or, z = 0.22) and 13 10-4 /Mpc (
0.77 103 Mpc ;or, z = 0.18). However, from Figure 6.27 that shows the periodogram
represented using the logarithmic scale, it is apparent that we can investigate
periodicities of frequency higher than f= 0.005 Mpc-1.

Figure 6.28 shows the

periodogram against the frequency range from 0.005 Mpc-1 (equivalent to 200 Mpc in
the comoving scale, which is of the order of the superclusters diameter) to 0.05 Mpc-1
54

(equivalent to 20 Mpc which is of the order of the voids diameter). The periodogram
shows that there could be a possible periodicity of f~0.006 Mpc-1. However, the peak is
close to the edge and its power is not much higher than the other peaks, so it is not
statistically significant.

55

56

Figure 6.18: Periodograms of the collective sample redshifts calculated using four different sampling rates.

57

Figure 6.19: The corresponding cumulative Periodograms of Figure 6.18s Periodograms.

58

Figure 6.20: Periodograms of the redshifts of sample 2 (the well-sampled region) calculated using four different sampling rates.

59

Figure 6.21: The corresponding cumulative Periodograms of Figure 6.20s periodograms.

60

Figure 6.22: Periodograms of sample 1 (upper panel) and sample 2 (lower panel) represented using logarithmic scale.

61

Figure 6.23: Periodograms of sample 1 for the high-frequency range.

62

Figure 6.24: Periodograms of sample 2 for the high-frequency range.

63

Figure 6.25: The periodogram of sample 3s data.

64

Figure 6.26: The cumulative Periodograms of sample 3s data.

65

Figure 6.27: Periodogram of sample 3 represented using logarithmic scale.

66

Figure 6.28: Periodogram of sample 3 for the high-Frequency range.

6.2. Visual explorations of clustering and connectivity


At the beginning of this research, one of the aims was to illustrate the data that
we collected in a 3D plots as a means of visually examining the degree of clustering and
filament connectivity. However, the illustrations generated by the software that we used
(MATLAB 2014b) served no use in the detection of the mass fluctuations and we were
unable obtain a suitable software that would enable us to carry out the goal within the
given period of time that was available to us.
Figures 6.29-30 illustrate the 3D distribution of the galaxies in the Cartesian
coordinates, converted from equatorial coordinates and galactic coordinates respectively.
These Figures show that the galaxies contained in the sample spread over the whole
ranges of the angular dimensions of the sky, although there is a region in which most of
the galaxies are contained. This can be further clarified using 2-D plots of the sky area
covered by the sample; such as those of Figures 6.28-29 where we can see clearly very
dense regions. From both illustrations (6.28 and 6.29) the Milky Way disk can be easily
noticed.
Although we gathered very large data set, the 3D figures illustrate clearly that the
sampling rates of galaxy are inhomogeneous and nonuniform and concentrate in the
northern celestial hemisphere. This is partly due to the presence of several observatories
in the North hemisphere of the Earth. Also, from the 2D plots it is obvious that the
surveys coverages vary in terms of the probing mechanism (i.e. the distributions of the
slices vs. pencil beams) between the two galactic hemispheres.
Although the 2D diagrams are very helpful to show the areas that are
highly/sparsely probed, these representations suffer from projection effects such that
some areas appear dark giving the impression that they are horizontally dense whereas
many of these galaxies are sparsely located along the radial direction. Therefore, both
2D and 3D representations should be concurrently used in the exploration and analysis
processes.

67

68

Figure 6.29: The 3-D distribution of galaxies covered by the sample, represented in the Cartesian coordinates
converted from equatorial coordinates

69

Figure 6.30: The 3-D distribution of galaxies covered by the sample, represented in the Cartesian coordinates converted from galactic
coordinates

70

, degree

Figure 6.31: The 2-D distribution of galaxies covered by the sample along the equatorial coordinates; ra and dec.

, degree

71

Figure 6.32: The 2-D distribution of galaxies covered by the sample along the galactic coordinates; land b.

, degree

Chapter 7: Summary and Conclusion


Redshift surveys have provided a wealth of information on the distribution of objects in
the universe and have become a primary source of information of too many
cosmological studies. However, each sky survey follows specific criteria for the sample
selection, based on the sky coverage/observing location or some photometric property.
In this study, one of the main aims is to construct a large sample by combining the data
from different redshift sky surveys to be used as an unbiased sample in the structure
studies of this project or any future work. However, such claim is not precisely valid as
the number of redshifts cross survey vary a lot (i. e. the SDSS huge size is likely to
influence our dataset).
The data of 2763064 objects have been obtained from combining ten redshift sky survey
catalogues, and after removing all of the duplicate and unreliable measurements. The
ten survey catalogues are: Sloan Digital Sky Survey (SDSS, 10th Release), Two Micron
All-Sky Survey (2MASS, the final data product), Two-degree-Field Galaxy Redshift
Survey (2dFGRS, the best spectroscopic observations of the final release), Six-degreeField Galaxy Redshift Survey (6dFGRS, Data Release 3), Galaxy And Mass Assembly
survey (GAMA, Data Release 2), VIMOS Public Extragalactic Redshift Survey
(VIPERS, PDR-1), The Deep Extragalactic Evolutionary Probe project (DEEP, Data
Release 4), VIMOS-VLT Deep Survey (VVDS, First Epoch sample), Team Keck
Redshift Survey (TKRS) and FORS Deep Field spectroscopic Survey (FDF).
This collective sample has been used to test the claim of periodicity in the radial
distribution of the celestial objects using a widely employed statistical technique for
detecting periodicity; the Periodogram Estimator. The periodicity test shows a possible
period of ~167 Mpc in the comoving distance scale and no periodicity in the redshift
scale. The inability to have 3D slicing utility limited our visual exploration of the
clustering and filament connectivity within the large dataset. We hope to revisit the
issue after much higher sampling rates can be achieved with the future telescopes such
as the James Webb Space Telescope (JWST) and the Thirty Meter Telescope (TMT).

72

Appendix A:
A MATLAB program for the periodogram calculations
function output=myperfun(z, Ts)
bincounts=histc(z,0:Ts:max(z));
Fs = 1/Ts;
% Ts is the sampling time, Fs is the Sampling rate
% Calculate time axis
tAxis =0:Ts:max(z);
y=bincounts;
L = length (y);

% Window Length of FFT

nfft = pow2(nextpow2(L));

% Transform length

y_HannWnd = y.*hanning(L);
Ydft_HannWnd = fft(y_HannWnd,nfft);
% at all frequencies except zero and the Nyquist (for a one-sided
% spectrum)
mYper = abs(Ydft_HannWnd).^2/L; % power of the DFT
mYper = mYper (1:nfft/2+1);
mYper (2:end-1) = 2* mYper(2:end-1);
f = Fs/2*linspace(0,1,nfft/2+1);
figure(1)
subplot(2,1,1)
bar(0:Ts:max(z),bincounts,'histc')
set(gca,'FontSize',20)
title(['Histogram of the Redshift Distribution in the Time Domain, bin size =
',num2str(Ts)] );
xlabel('z');
ylabel('Counts');
grid on
subplot(2,1,2)
plot(f,2*mYper,'.-'); % we need 2 because Hanning Wnd Amplitude Correction Factor
=2
73

set(gca,'FontSize',20)
grid on
title(['Priodogram with Hann Wnd, Sampling rate = ', num2str(Fs)]);
xlabel('Frequency (Hz)');
ylabel('Power');
hold on
[positions, extrema]=findextrema(f,2*mYper);
% plot(positions,extrema,'or')
peak_or_valley=diff(extrema);
npeak=find(peak_or_valley>0)+1; peaks=extrema(npeak);
nvalley=find(peak_or_valley<0)+1; valleys=extrema(nvalley);
plot(positions(npeak),peaks,'or')
output=positions(npeak)';
hold off
%---------- Comulative Periodogram ---------%
Comper=[];
for k=1:nfft/2+1
comper=sum(mYper(1:k))/sum(mYper(1:nfft/2+1));
Comper=[Comper, comper];
end
figure(2)
plot(f,Comper)
set(gca,'FontSize',20)
grid on
title(['Sampling Time = ',num2str(Ts)] );
xlabel('Frequency')
ylabel('Comulative Periodogram')
zoom xon
end

74

References
[1] (1994). Encyclopedia of Time. In S. L. Macey (Ed.), Garland reference library od social
science. USA: Taylor & Francis.
[2] (1994). Relativity in General. In J. D. Alonso, & M. L. Paramo (Eds.), Proceedings of the
Relativity Meeting '93, Salas, Asturias, (Spain), September 7-10, 1993. Atlantica Sguier
Frontires.
[3] Cosmic Microwave Background seen by Planck. (2013, December 17). (ESA, Planck
Collaboration) Retrieved October 24, 2014, from ESA/Planck:
http://sci.esa.int/planck/51553-cosmic-microwave-background-seen-by-planck/
[4] (2014). Retrieved June 14, 2014, from Sloan Digital Sky Surveys:
http://www.sdss.org/sdss-surveys/
[5] Ade , P., Aghanim, N., Alves, M. I., Armitage-Caplan, C., Arnaud, M., Ashdown, M., et al.
(2013, 3). Planck 2013 results. I. Overview of products and scientific results.
[6] Angelo, J. A. (2009). Encyclopedia of Space and Astronomy: Facts on File science library.
New York, USA: Infobase Publishing.
[7] Appenzeller, I. (2009). High-Redshift Galaxies, Light from the Early Universe (1st ed.).
London: Springer Dordrecht Heidelberg.
[8] Baldry, I. K., Robotham, A. S., Hill, D. T., Driver, S. P., Liske, J., Norberg, P., et al. (2010,
May). Galaxy And Mass Assembly (GAMA): the input catalogue and star-galaxy
separation. Monthly Notices of the Royal Astronomical Society, 404(1), 86-100.
[9] Bartusiak, M. (2011, 9). The Day We Found the Universe: The Little-Known History of
How We Came to Understand the Expanding Universe. (Joseph B. Jensen, James G.
Manning, & Michael G., Eds.) 25.
[10]
Baum, W. A. (1962). Photoelectric Magnitudes and Red-Shifts. (G. C. McVittie,
Ed.) International Astronomical Union Symposium, 15(Problems of Extra-Galactic
Research,), 390.
[11]

Belenkiy, A. (2014, 3). Discovery of Hubble's Law: an Example of Type III Error.

[12]
Bennett, C., Larson, D., Weiland, J., Jarosik, N., Hinshaw, G., Odegard, N., et al.
(2013, 6 2013). Nine-Year Wilkinson Microwave Anisotropy Probe (WMAP)
Observations: Final Maps and Results.
[13]
Bodenheimer, P. (2011). Principles of Star Formation. Springer Science &
Business Media.

75

[14]
Chatfield, C. (2004). The Analysis of Time Series: An Introduction (6th ed.). Boca
Raton, Florida, USA: Chapman & Hall/CRC Texts in Statistical Science.
[15]
Colless, M., Peterson, B. A., Jackson, C., Peacock, J. A., Cole, S., Norberg, P., et
al. (2003, June). The 2dF Galaxy Redshift Survey: Final Data Release.
[16]
Cosmic Web and Formation of Galaxies . (n.d.). (Jacobs University Bremen)
Retrieved 10 30, 2014, from Modern Cosmology - From the Big Bang to Cosmic Web :
http://www.lsw.uni-heidelberg.de/users/mcamenzi/Week_7.html
[17]
Cosmological Redshift. (n.d.). (S. U. Technology, Producer) Retrieved from
COSMOS - The SAO Encyclopedia of Astronomy:
http://astronomy.swin.edu.au/cosmos/c/cosmological+redshift
[18]
Dardo, M. (2004). Nobel Laureates and Twentieth-Century Physics (Illustrated
ed.). Cambridge, UK: Cambridge University Press.
[19]
Davis, M., Faber, S. M., Newman, J., Phillips, A. C., Ellis, R. S., Steidel, C. C., et al.
(2003, February). Science Objectives and Early Results of the DEEP2 Redshift Survey. (P.
Guhathakurta, Ed.) Discoveries and Research Prospects from 6- to 10-Meter-Class
Telescopes II, 4834, 161-172.
[20]
Diggle, P. J., & Fisher, N. I. (1991). Nonparametric Comparison of Cumulative
Periodograms. Journal of the Royal Statistical Society. Series C (Applied Statistics),
40(3), 423-434.
[21]
Ebrahimi, E., & Riazi, N. (2012, 5). Can Inflation, Dark Matter and Dark Energy
Be Described in Terms of a Single, Real Scalar Field? 51(5).
[22]
Feigelson, E., & Babu, G. (1996). Astrostatistics (1st ed.). London, UK: Chapman
& Hall.
[23]
Franzetti, P. (n.d.). Retrieved June 12, 2014, from VIMOS Public Extragalactic
Redshift Survey (VIPERS): http://vipers.inaf.it/
[24]
Garner, R. (Ed.). (2012, 5 31). NASA's Hubble Shows Milky Way is Destined for
Head-On Collision. Retrieved 7 22, 2014, from NASA:
http://www.nasa.gov/mission_pages/hubble/science/milky-way-collide.html
[25]
Gegory, S., & Laird, A. (1978, June 15). The COMA/A1367 Supercluster and its
Environs. The Astrophysics Journal, 222, 784-799.
[26]
Gillessen, S., Eisenhauer, F., Trippe, S., Alexander, T., Genzel, R., Martins, F., et
al. (2009, February). Monitoring Stellar Orbits Around the Massive Black Hole in the
Galactic Center. 692(2), 1075-1109.

76

[27]
Giovanelli, R., & Haynes, M. P. (1988). Extragalactic Neutral Hydrogen. In G.
Verschuur, & K. Kellermann (Eds.), Galactic and Extragalactic Radio Astronomy (2nd
ed.).
[28]
Godart, O., & Heller, M. (1985). Cosmology of Lemare. In History of astronomy
series (Vol. 3). Parchart Publishing House.
[29]
Gregory, S. A., Thompspn, L. A., & Tifft, W. G. (1981, January 15). The Perseus
Supercluster. The Astrophysics Journal, 243, 411-426.
[30]
Gribbin, J. (1998). In the Search of the Big Bang, The life and Death of the
Universe (2nd ed.). London: Penguin Books.
[31]
Hathaway, D. H. (2014, July 9). Sun Facts. Retrieved July 12, 2014, from
National Aeronautics and Space Administration, NASA:
http://webcache.googleusercontent.com/search?q=cache:http://solarscience.msfc.nas
a.gov/
[32]
Hawkins, E., Maddox, S. J., & Merrifield, M. R. (2002). No Periodicities in 2dF
Redshit Survey Data. Monthly Notice of the Royal Astronomical Society.
[33]
Hearn, G., & Metcalfe, A. (1995). Spectral Analysis in Engineering: Concepts and
Case Studies. Oxsford, UK: Butterworth-Heinemann.
[34]
Huchra, J. P., Macri, L. M., Masters, K. L., Jarrett, T. H., Berlind, P., Calkins, M.,
et al. (2012, April). The 2MASS Redshift SurveyDescription and Data Release. The
Astrophysical Journal Supplement, 199(2).
[35]
Huchra, J. (n.d.). The CfA Redshift Survey. Retrieved June 13, 2014, from Center
For Astrophysics: https://www.cfa.harvard.edu/~dfabricant/huchra/zcat/
[36]
Jones, D. H., Read, M. A., Saunders, W., Colless, M., Jarrett, T., Parker, Q. A., et
al. (2009, October). The 6dF Galaxy Survey: final redshift release (DR3) and southern
large-scale structures. Monthly Notices of the Royal Astronomical Society, 399(2), 683698.
[37]
Jones, M. H., & Lambourne, R. J. (Eds.). (2004). An Introduction to Galaxies and
Cosmology. Cambridge, UK: Cambridge University Press.
[38]
Kambic, B. (2009). Viewing the Constellations with Binoculars: 250+ Wonderful
Sky Objects to See and Explore. New York, USA: Springer Science & Business Media.
[39]
Kang, E. W. (2008). Radar System Analysis, Design, and Simulation. Norwood,
USA: Artech House.

77

[40]
Lahav, O., & Suto, Y. (2004, August 4). Measuring our universe from galaxy
redshift surveys.
[41]
L'Annunziatea, M. F. (2007). Radioactivity: Introduction and History (1st ed.).
Oxford, UK: Elsevier.
[42]
Le Fvre, O., Vettolani, G., Garilli, B., Tresse, L., Bottini, D., Le Brun, V., et al.
(2005, September). The VIMOS VLT deep survey. First epoch VVDS-deep survey: 11 564
spectra with 17.5 IAB 24, and the redshift distribution over 0 z 5. Astronomy and
Astrophysics, 439(3), 845-862.
[43]

Li, X., Qin, H., & Zhang, T. (2014, 4). Fluctuation of Hubble Parameter.

[44]
Martine, V., & Saar, E. (2002). Statistics of the Galaxy Distribution (1st ed.).
Boca Raton: Champman & Hall/CRC.
[45]
Michel-Marie, D., & Elena, D. (2006). Dictionary of Distances (Revised ed.).
Elsevier.
[46]
Napier, W. M., & Guthrie, B. N. (1997). Quatized Redshifts. Jornal of
Astrophysics and Astronomy.
[47]
Noll, S., Mehlert, D., Appenzeller, I., Bender, R., Bhm, A., Gabasch, A., et al.
(2004, May). The FORS Deep Field spectroscopic survey. Astronomy and Astrophysics,
418, 885-906.
[48]
O'Callaghan, J. (2012, May 29). Expansion of the universe. (D. Butt, S. Boyd, M.
Peron, A. Asadi, Producers, & Imagine Publishing Ltd) Retrieved from Space Answers:
http://www.spaceanswers.com/deep-space/expansion-of-the-universe/
[49]
O'Raifeartaigh, C. (2013, 4). The Contribution of V. M. Slipher to the Discovery
of the Expanding Universe. 471(Origins of the Expanding Universe: 1912-1932), 49.
[50]
Rhee, J., Zwaan, M. A., Briggs, F. H., Chengalur, J. N., Lah, P., Oosterloo, T., et al.
(2013, August). Neutral atomic hydrogen (H I) gas evolution in field galaxies at z ~ 0.1
and 0.2.
[51]
Robinson, M. R. (1981). Cosmology (2nd ed.). New York: Oxford University
Press.
[52]
Smith, T. P. (2013). How Big is Big and How Small is Small: The Sizes of
Everything and Why. USA: Oxford University Press.
[53]
Soter, S., & Tyson, N. D. (Eds.). (2000, 8 28). Georges Lematre, Father of the Big
Bang. Retrieved 2014, from American Museum of Natural History:

78

http://www.amnh.org/education/resources/rfl/web/essaybooks/cosmic/p_lemaitre.ht
ml
[54]

Straumann, N. (2002, 8). The history of the cosmological constant.

[55]
Strauss, M. A., & Willick, J. A. (1995). Redshift Surveys: Setting the Quantitaitve
Groundwork. 261(5-6), pp. 271-431.
[56]
Surendran, D. (2011, November 29). redshift2dist.m. (Instituto de Fsica y
Matemtica) Retrieved June 16, 2014, from
http://www.ifm.umich.mx/~villasen/Hugo_Tesis_Nov_2011/redshift2dist.m
[57]
Takahashi, F., Kondo, T., Takahashi, Y., & Koyama, Y. (2000). Wave summit
course: Very Long Baseline Interferometer. Japan: IOS Press.
[58]
Tang, S. M., & Zhang, S. N. (2005). Critical Examinations of QSO Redshift
Periodicities and Associations with Galaxies in Sloan Digital Sky Survey Data. The
Astrophysics Journal, 633(1), 41-51.
[59]

Thompson, L. A. (2011, 8). Vesto Slipher and the First Galaxy Redshifts.

[60]
Tifft, W. G. (1979). Periodicity in the redshift intervales for double galaxies. The
Astrophysical Journal, 70-74.
[61]
Trimble, V. (1996, 12). H_0: The Incredible Shrinking Constant, 1925-1975. 108,
1073-1082.
[62]
Trimble, V. (2000). The Ratios of Small Whole Numbers: Misadventures in
Astronomical Quantization. In Rene Donaldson , & Bill Kirk (Eds.), Special Quantum
Century (Vol. 30). DIANE Publishing.
[63]
Tropp, E. A., Frenkel, V. y., & Chernin, A. D. (2006). Alexander A Friedmann: The
Man Who Made The Universe Expand. Camobridge: Campbridge University Press.
[64]
Vio, R., Diaz-Trigo, M., & Andreani, P. (2013, February). periodogram, Irregular
time series in astronomy and the use of the Lomb-Scargle. 1, 5016.
[65]
Wirth, G. D., Willmer, C. N., Amico, P., Chaffee, F. H., Goodrich, R. W., Kwok, S.,
et al. (2004, June). The Team Keck Treasury Redshift Survey of the GOODS-North Field.
The Astronomical Journal, 127(6), 3121-3136.
[66]
Woolfson, M. M. (2013). Time, Space, Stars & Man: The Story of the Big Bang
(2nd ed.). Toh Tuck Link, Singapore: World Scientific.
[67]
York, D. G., Adelman, J., Anderson, J. E., Anderson, S. F., Annis, J., Bahcall, N. A.,
et al. (2000, 9). The Sloan Digital Sky Survey: Technical Summary. The Astronomical
Journal, 120(3), 1579-1587.
79

80

S-ar putea să vă placă și