CHAPTER 1

INTRODUCTION
1.1 Importance of atmospherical vapour for the Climate System
Water vapour is continually athletics through the atmosphere, evaporating from the surface, compressing to
create clouds and afterwards returning to the world as precipitation. Heat from the sun evaporates water and
this heat is free into the air once the water condenses into clouds and once it precipitates. The evaporation
condensation cycle is a crucial mechanism for transferring heat from the Earth's surface to its atmosphere
and in moving heat round the earth. thus atmospherical vapour is very important for the climate system and a
key to grasp the hydrological cycle.
Trends
Recent world estimates of long tropospheric water vapour changes show a rise in precipitate water
throughout the amount from 2005 to 2015. Largest trends will be ascertained within the tropics with a rise of
concerning thirteen per decade. associate degree analysis of vapour trends higher than North America
supported radiosonde measurements from 1993 to 2013 reveals will increase in precipitate table water over
all regions except northern and jap Canada wherever it fell slightly. Regions of wetness increase ar related to
regions of rising temperatures over identical amount, and also the regions of weakened wetness ar related to
falling temperatures. One major hindrance for associate degree improved information of the role of vapour
within the climate system is joined to data-based gaps that presently exist. Most studies primarily have
confidence radiosonde information. they need a fairly smart vertical, however a poor horizontal resolution.

1.2 The Role of water vapour in GPS geophysics and Navigation

1.2.1 NAVSTAR GPS
NAVSTAR GPS is funded and controlled by the U. S. Department of Defense (DOD) and consists of 3
segments, the space, the management and also the user section. The house and also the user section ar
shortly explained here.
The GPS constellation consists of a minimum of twenty one satellites organized in vi orbitalnplanes with 55
inclination associate degreed an altitude of twenty,200 klick higher than the Earth's surface. The orbital
amount is concerning twelve hours, thus a GPS satellite is incessantly visible higher than the horizon for
concerning five hours. many varieties of GPS satellites are launched to date. Most of the satellites presently
in use ar Block II and Block IIA satellites. The latter ar simply alittle modification of the initial style. Block
II satellites have a weight of 930 kilo and a velocity of four km/s.
They transmit signals victimization 2 frequencies, particularly L1 (fL2 = 1575.42 MHz) and L2 (fL2 =
1227.60 MHz) and receive signals on a frequency of 1783.74 MHz. Block II satellites carry four atomic
clocks - two atomic number 37 and a couple of Cs clocks - and have a design-lifetime of seven.5 years. The
renewal satellites of kind Block IIR have the next lifespan of ten years.
The user section contains the GPS receivers required to decipher the transmitted signals. For tropospheric
delay estimation , high-quality two-frequency receivers ar necessary so as to eliminate ionospheric
propagation delays properly.
1

1.2.2 Tropospheric Delays and GPS

Precise GPS applications ar difficult by atmospherical effects, particularly propagation delays within the
region and layer. As so much because the region cares, it's potential to compensate the primary order impact
given measurements on 2 totally different frequencies. For the delay caused by the layer, no dispersion
effects ar gift and elimination isn't potential. The tropospheric delay will be separated into a
hydrostatic and a wet part. The hydrostatic part in celestial point direction is named ZHD (zenith hydrostatic
delay). It will be exactly determined by surface pressure measurements. The ZWD (zenith wet delay),
however, can not be sufficiently modelled by surface measurements thanks to the irregular distribution of
water vapour within the atmosphere. Next chapter can affect the main points of tropospheric delay
modelling.
The ZHD amounts to concerning two.3 m, whereas the ZWD is simply within the vary of zero.15 m in world
average. though it's a lot of smaller than the hydrostatic part, the uncertainties in we tend tot tropospheric
delay modeling do place an excellent burden on high exactitude GPS applications if we recall that carrier
section measurements themselves have associate degree accuracy of a couple of milli meters and bound
receiver makers even claim that they need reached noise levels within the sub-milli-meter vary.
Among several examples in geophysics and navigation, NAITO [1998] reports a few Japanese network that
has been established for deformation analysis and earth-quake detection. The analysis of the measurements
showed drifts within the coordinate solutions that apparently weren't thanks to plate motions, however
associated with associate degree light compensation of tropospheric delays. extra estimation of wet delays
inside the routine analysis ought to improve the case.

1.3 vapour perceptive Systems

A variety of platforms and sensors is out there to live atmospherical water vapour. Each system has bound
blessings and drawbacks. the subsequent summary is based on MOCKLER and Gabor. Figure 1-1 shows a
number of these sensors and their characteristics ar summarized in Table 1-1.
1.3.1 Description of hand-picked Sensors
Ground-based direct sensors, weather balloons and craft were the sole tools offered to live precipitable
vapour within the past. The radiosonde has been one amongst the foremost vital devices and can actually still
be a valuable vapour detector within the future. Radiosondes ar balloon-borne instruments with radio
transmittal capabilities. They contain instruments capable of creating direct unaltered measurements of air
temperature, humidness and pressure with height, usually to altitudes of roughly thirty klick. These
ascertained information ar transmitted instantly to the bottom station by a sender. Ground-based radio
direction finding antenna instrumentality tracks the motion of the radiosonde throughout its ascent through
the 40 On Ground-based GPS Tropospheric Delay Estimation air. The recorded elevation and AZ info ar
reborn to wind speed and direction at numerous levels by triangulation techniques.
Today, more remote sensors capable of taking vapour measurements over giant aras are offered. Substantial
progress has been created victimization satellite observations to get total column water vapour and lowresolution vertical profiles from infrared and microwave sensors, however these satellite observations don't
give information underneath all atmospheric condition nor specially surfaces. typically speaking, they live
2

absorption lines within the radiation from the new background provided by the world. Consequently, clouds
could cause issues in addition because the surface of land plenty. Best results ar obtained over oceanic
regions.
Hygrometers and LIDARs will be helpful for elaborate native studies. measuring instrument is associate
degree form for lightweight detection and locomote, a vigorous remote sensing technique that operates in an
exceedingly similar manner as echo sounder systems: A pulse of optical maser lightweight is emitted into the
sky and also the quantity of come thanks to break up from the atmosphere is measured versus time. With
information of the speed of sunshine, the time is reborn into altitude. The number of photons counted for
every altitude bin is proportional to the atmospherical density.

Figure 1-1 - summary of water vapour measure platforms and sensors following CARTER Water vapor sensors ar
carried on a spread of platforms as well as ground stations, weather balloons, craft and satellites. NAVSTAR GPS
consists of a minimum of twenty one MEO satellites at associate degree orbit altitude of twenty,000 km.
atmospherical observance with GPS is feasible in 2 ways: by searching the atmosphere with facilitate of GPS
receivers on LEO (low-earth orbiting) satellites, the supposed radio break technique, or by victimization networks
of ground-based GPS receivers. in keeping with GABOR [1997], remotely piloted vehicles have nevertheless to
return ancient, however supply a doubtless long loiter time and high altitude ceiling.

1.4 Objectives and Structure of this Thesis

The outstanding problems in vapour analysis for meteorology ar printed by MOCKLER and divided into
theoretical, data-based and climate modelling problems. To summarize the primary 2 classes, an absence of
data as so much because the role of vapour in influencing the radiation budget of the world will be declared.
identical is true for the processes decisive the distribution of vapour and its changes over time. Retrievals of
vapour shall be improved and extended with special stress on long, continuous and world observations to
assist within the analytic thinking.
In GPS geophysics and navigation, the first goal in modelling the tropospheric propagation delay is to cut
back the uncertainties associated with the wet part or, equivalently, to water vapour. For permanent GPS
3

arrays, the extra estimation of celestial point wet delays reduces the impact of tropospheric modelling errors
and provides valuable estimates of total atmospherical column water vapour for climate models and
meteorology. for several different GPS applications, tropospheric delay estimation isn't potential or
problematic. A typical example for this cluster are kinematic applications that commonly don't enable the
determination of such extra parameters and are thus subject to improvement by precise tropospheric
corrections equipped by GPS permanent arrays and/or numerical weather models.
1.4.1 Objectives of this Thesis
In order to take advantage of the potential of GPS receivers as water vapour sensors, a tropospheric analysis
system is to be developed that's able to exactly filter celestial point wet delays and integrated water vapour
from GPS section measurements. This platform independent process package ought to be operable in semiautomatic and totally automatic mode as well as non-interactive information writing. computer file in
commonplace formats should be supported and also the analysis results got to be hold on in normally united
commonplace formats, too. Special stress is to be arranged on the responsibility of the results, and also the
network performance is to be mechanically analyzed and, if necessary, to be altered (i. e. exclusion of
suspicious receivers) and re-analyzed consequently. The computer code shall be able to analyze multi-station
networks with long baselines (more than one thousand km) with enough accuracy. The temporal resolution
is to be chosen adequately and shall meet potential desires for high-frequency information.
The overall goal is to estimate wet delays and integrated water vapour whereas the quality output of
complete GPS information analysis will solely encompass neutral (or total) delays. this suggests that extra
meteorologic information ar required like surface pressure for modeling the hydrostatic part and
temperature for the conversion of wet delays into precipitable/integrated water vapour. Therefore, the
analysis system should give all necessary interfaces to access these extra information. As a matter of reality,
the supply of in place meteorologic measurements at GPS monitor stations is extremely restricted. As a
consequence, algorithms got to be developed to extract the required info from numerical weather models.
thanks to the world size of GPS following web works just like the IGS net, weather models of worldwide
extend are favorable to accomplish this task.
The efforts to extract tropospheric info from numerical weather models shall not be restricted to the support
of the GPS information analysis with reference to meteorology and meteorology, however shall even be
assessed in terms of their relevance to GPS navigation and geophysics. An open, memory-efficient and
gridded format shall be developed that carries all information analyzed within the three-D numerical weather
models that ar required to work out hydrostatic , wet and neutral delays at anyplace on the world. These
gridded tropospheric correction information shall be suited to kinematic GPS process and may conjointly
function a carrier for gridded integrated water vapour information. correct ways for horizontal interpolation
and vertical reduction of the delays are to be developed and valid.
Finally, the on an irregular basis distributed tropospheric delays calculable at the GPS sites shall be
combined with the gridded information of the tropospheric correction files that are derived from numerical
weather models. during this manner, the GPS and also the less correct NWM delays are liquified along and
also the outcome are associate degree improved, additional correct answer for the gridded information sets.
appropriate algorithms for this mix shall be
investigated with special concentrate on random improvement.

All algorithms ar to be documented and mentioned. This doesn't solely discuss with the special tasks of GPS
tropospheric delay estimation, however conjointly includes a full summary of the components which will be
referred to as "conventional" GPS processing, i. e. those algorithms that ar essential for GPS information
analysis, albeit not primarily associated with GPS tropospheric delay estimation. The dominant error sources
shall be analyzed and their impact is to be evaluated with facilitate of sensible experiments. All 3
cornerstones, particularly (I) the filtering of celestial point wet delays and integrated water vapour from GPS
measurements, (II) the meteorologic information extraction from numerical weather fields and its application
to GPS meteorology, and (III) the suitableness of the gridded tropospheric correction information from
weather fields in addition because the combined GPS/NWM answer fields ought to be valid.

Figure 1-2 - Structure of this thesis. Chapters two to four will be thought of because the theoretical
half describing the modelling ways and algorithms of GPS tropospheric delay estimation and
information extraction from numerical weather fields. sensible results ar bestowed in chapters five to
seven and a summary is given in chapter eight.

Chapter -2
Principles of GPS processing
Tropospheric delay estimation needs careful modelling of the GPS measurements. For this reason, the
principles of GPS processing ar printed here within the initial place before the matter of tropospheric
propagation delays is self-addressed within the chapter three well. Note that the subsequent sections aren't
meant as a general introduction into GPS process, however concentrate on the ways enforced within the
tropospheric analysis computer code Trop AC on that the results bestowed in chapters five, vi and seven are
based mostly. For a general description, the interested reader could discuss with SEEBER and
HOFFMANN-WELLENHOF et al. [1993], as an example.

2.1 process summary

The main process steps will be divided into pre-processing, network filtering and post-filter process. These 3
stages ar shortly mentioned below. Links to the corresponding sections are given.
2.1.1 Pre-Processing
At the pre-processing stage, the required information ar scan, checked and corrected. The measurements like
code ranges and carrier section observations ar scan from RINEX files. Carrier phases need some extra
corrections, e. g. the synchronization downside has got to be self-addressed if (single or) double variations
shall be shaped, the orientation downside will be relevant for extended baselines, elevation-dependent
antenna section center corrections should be applied. Finally, ionospheric and tropospheric delays ar
foreseen. Moreover, typical issues like cycle slip detection and repair in addition as multipath detection
conjointly belong to the tasks of the pre-processor.

Figure 2-1 - Major modules of the Trop AC permanent array filter computer code and their
interrelations. The pre-processor for the GPS information is named PAF_PREP and prepares all
required information for Kalman filtering. Network filtering is allotted with the modules PAF_FILT,
PAF_TROP or PAF_RANG. Program PAF_MEO is interface to the numerical weather models and
PAF_COMB is ready to combine often gridded tropospheric information sets from the weather fields
and also the on an irregular basis distributed GPS estimates.

2.1.2 Network Filtering

The pre-processed information are filtered for the unknowns, particularly ground station coordinates (and in
special cases conjointly satellite orbit biases), celestial point wet delays and ambiguities. For this purpose, a
Kalman filter is employed. though pre-processing ought to have removed suspicious information, the filter
itself is capable of blunder detection and should even perform a post-fit residual analysis and tell on the
previous answer if necessary. As throughout the pre-processing steps, the network is checked for poorly
playacting sites another time ensuing in an exceedingly new filter run if such stations are found.
2.1.3 Post-Filter process
At the post-filter stage, bound computations are allotted that don't belong on to the Kalman filter method, e.
g. the derivation of mean ground station coordinate estimates, the information transformation of networks in
addition because the creation of free network answers by similarity transformation and also the method of
melting network partitions to associate degree overall solution file.
Certainly, troposphere-related jobs just like the conversion of celestial point wet delays into integrated water
vapour belong to the present process stage, too.

Name
Module

TROPAC

PAF_ORBI

PAF_MEO

PAF_COMB

PAF_PREP

PAF_FILT

of

Purpose

Guidance program; user interface to simplify access to the program system generation
of batch-files, scheduling of processes, visualization.
TropAC normally uses precise orbits from the IGS analysis centers which are provided
in the SP3-format defined by the National Geodetic Survey (NGS, USA), seeREMO-ND Y [1989, 1991] for details on this format. Broadcast orbits are less accurate and
given in a different format (RINEX NAV files). This module allows to created SP3files from one or more RINEX navigation files.
The meteorological module of TropAC. It allows to extract meteorological data from
numerical weather models provided in GRIB-format. Several data can be extracted like
surface pressure, temperature and humidity. TROPEX files can be generated and raytracing can be performed.
The tropospheric combination module. It allows to combine GPS-derived tropospheric
delays with those integrated in numerical weather fields.
The GPS pre-processor for static networks. This module prepares all needed data for
the filter process: GPS measurements are read and filtered, orbits are interpolated,
double differences are formed and synchronized, cycle slips are detected and repaired,
tropospheric delays are predicted and, finally, a binary network file is created
containing all necessary information for the filter engine.
The Kalman filter engine based on double-difference phase measurements. The state
vector of this filter consists of 3 coordinate components for each site and - optionally of one zenith wet delay per site modeled as random walk stochastic process. Moreover,
orbit relaxation can be applied. The ambiguity parameters are dynamically allocated in
the state vector.
This module equals PAF_FILT except for the fact that it does not estimate
7

PAF_TROP

coordinates. This reduces the number of unknowns and speeds up

processing.

2.2 Observations and Observation Equations

GPS observations4 principally comprise pseudo-range and carrier section measurements. section
measurements have a amplitude of a couple of millimeters and are terribly precise compared to code ranges
that are solely correct to a couple of meters or some decimetres at their best. For this reason, carrier phases
are the first and most vital kind of observation for high-precision parameter estimation. the aim of pseudoranges is primarily associated with the pre-processing stages. In distinction to carrier phases, code ranges
aren't ambiguous and might thus be simply applied to sight section breaks, as an example, and serve for
several different functions as are mentioned within the following sections.
2.2.1 Pseudo-Ranges
Pseudo-ranges are non-ambiguous measurements. though they need the next amplitude than section
measurements, they will serve an honest job for the detection of multipath, the paradox resolution method
and also the derivation of approximate station coordinates. Moreover, code ranges play a key-role for the
determination of the signal transmission epoch and facilitate to synchronize the baselines5. Following
WBBENA simplified observation equation for radio radiation L1 will be expressed as:

PRiA[L1] S A[L1] c (tA ti ) Si

A[ION,L1]

SA[TROP] PR

Si A: geometric distance between receiver antenna A and satellite i

C : speed of light (in vacuum, c = 299 792 458 m/s)
dtA : receiver clock error
dti : satellite clock error
dSiA [ION]: ionospheric propagation delay
dSiA[TROP]: tropospheric propagation delay
ePR: noise term
2.2.2 Carrier Phase Measurements
Precise positioning and filtering of tropospheric delays is carried out using carrier phase measurements. The
observation is the so-called carrier beat phase.

iA (tOR REC (tOR) REF (tOR) i tOT REC t(OR)

tOR: time of signal receipt
tOT: time of signal transmission
iA: carrier phase measurement, carrier beat phase
REC: received carrier phase from satellite i at receipt time tOR
i: carrier phase of satellite i at signal transmission time tOT
REF: phase of reference signal generated by the receiver A at receipt time tOR
2.2.2.1 Double Differences
8

The original carrier phases are not used for filtering because these measurements are corrupted by receiver
and satellite clock errors. Instead, double differences are derived: In a first step, single differences6 are
formed, i. e. phase differences between two ground stations A and B which both have satellite i in view. In a
second step, the same is done with those measurements to satellite j. The difference between the two single
differences is computed and yields the doubly differenced phase observation

ijAB[L1] j B[L1] iA[L1] iB[L1] iA[L1]

2.2.2.2 Synchronization Problem
Due to the presence of receiver clock errors, care must be taken to refer the measurements to the correct
epoch. Following the definitions given by WBBENA, the receiver time t at the measurement epoch
deviates from the true GPS time t by the receiver clock error tA in the following way:

tRtR t tR tRtA

tOR: true GPS time of signal receipt at receiver site A

t'OR: receiver time of signal receipt
tA: synchronization error, receiver clock error
2.2.2.3 Antenna Orientation downside

Circularly polarized magnetic attraction waves are transmitted from the GPS satellite's antenna to the
receiving antenna. Consequently, the ascertained carrier section measurements rely upon the orientation of
the receiver and transmitter antennas and also the direction of the road of sight. If the orientation between the
transmittal and also the receiving antenna changes, a amendment will be measured within the ascertained
section. For static networks, such changes in orientation are caused by the moving GPS satellites. WU et al.
[1993] incontestable that the correction thanks to antenna orientation is negligible for brief baselines,
however could reach a magnitude of up to four cm for baselines as long as 4000 kms.
The formula given here permits a general description of the matter and might even be applied for kinematic
GPS process. However, angle and AZ info is required in such cases as a result of the orientation of the
antenna dipoles should be renowned and should actually vary within the case of moving antennas. The
effective dipole of the receiving antenna is found to be

dA eX eiA (eiA eX ) eiA eY

dA: effective dipole vector of the receiving antenna at site A referring to satellite i
eX: ECEF unit vector in direction of the x-dipole element of the receiving antenna
eY: ECEF unit vector in direction of the y-dipole element of the receiving antenna
eiA: ECEF unit vector from receiving to transmitting antenna.
where all vectors ar expressed within the world, philosopher ECEF system. For static networks with the
bottom antennas being properly aligned to northward direction, the native level unit vectors of the 2 antenna
dipole parts ar outlined as

ex =

1
0

eY = - 1
9

ex: native level orientation vector of x-dipole component

eX: ECEF orientation vector of x-dipole component
2.2.4.1 initial Order impact :

Knowledge of the negatron content is enough so as to model the primary order ionospheric impact. In most
cases, solely the primary order delay is taken into consideration as a result of it's out and away the foremost
dominant a part of the ionospheric propagation delay.

nION = 1+ C2 /f2 = 1+C.Ne

Ne: negatron density in [electrons/m]
C: constant; C = forty.28 [m/s]
Now, the ionospheric propagation delay in celestial point direction will be integrated victimization the
negatron density
Ne. For carrier section measurements this correction in celestial point direction is

Z=0
=
ION

Cf 22
hA

C
f2

N
hA

(h).dh =

C
f2

.VTEC

10

CHAPTER -3
Modelling and Estimating Tropospheric Propagation Delays
The primary purpose of the TropAC analysis system is to estimate wet tropospheric delays which will be
reborn into integrated vapour and thereby function a valuable input into numerical weather and climate
models. For this reason, a whole chapter is dedicated to tropospheric delay modelling and estimation.
3.1 temporary summary of the Lower Atmosphere
The atmosphere may be a mixture of dry gases that primarily contribute to the hydrostatic delay and water
vapour that is answerable for the wet delay. SPILKER states that the dry atmosphere is uniform and
uncomplicat- -ed in modelling whereas the wet half is erratically distributed. The vertical profiles of the
foremost vital meteorologic quantities ar planned in Figure 3-2 for associate degree at random chosen day at
IGS monitor station Oberpfaf fenhofen .
The following layers ar to be distinguished: The layer ranges from water level ( zero m) to a height of
concerning twelve klick and is characterised by a comparatively linear temperature decrease. The layer may
be a little
boundary layer between twelve and sixteen klick wherever the temperature remains some constant at tier of
-60 to -80 C and within the layer (16 to fifty km), a slow temperature increase happens. As so much because
the wet
component cares, the lower layer is of major interest whereas the hydrostatic part is influenced up to the
stratopause. SPILKER mentions that concerning one quarter of the entire delay is caused by gases higher
than the layer. Most of the vapour contents, however, is targeted at a height right below four klick and higher
than twelve klick, nearly no additional vapour is gift.

11

Figure 3-1 - Schematic of the tropospheric and stratospheric

layers and the tropopause after MOCKLER [1995]. The mean
vertical distribution of temperature and water vapor mixing
ratio (logarithmic scale) are shown.

3.1.1 Pressure
The sea level pressure is concerning 1013 hPa in average, a worth that's utilized in most traditional
atmosphere models. Figure 3-2 shows that the pressure decreases exponentially with increasing altitude. The
layer height is reached at a pressure between three hundred hPa at the poles and seventy hPa at the equator
and a worth of roughly one hPa will be declared at the stratopause height.
3.1.2 Temperature
The temperature shows a linear decrease up to the layer, however note that this linear trend will be
significantly disturbed within the initial few hundred meters higher than the surface thanks to inversion
layers. The supposed temperature lapse rate is within the vary of -5 to -7 K/km below the layer height. At the
layer itself, the temperature remains some constant and slowly will increase within the layer.
3.1.3 water vapour
The ratio diagram in Figure 3-2 indicates that not solely the horizontal distribution, however conjointly the
vertical distribution of water vapour within the layer can not be expected to be uniform. the explanation for
this is often joined to the fast turnover of water within the air in addition on the variation of temperature with
height and location1, see MOCKLER .
Although there ar appreciable variations, Figure 3-1 conjointly implies that there ar bound trends: water
vapour decreases quickly with height because the atmosphere gets colder. Nearly [*fr1] the entire water
within the air is found between water level and concerning one.5 klick higher than water level. but five-6%
12

of the water is gift higher than 5 klick, and fewer than a hundred and twenty fifth is found within the layer.
ratio shown in Figure 3-2 conjointly tends to decrease with height, from a mean price of concerning 60-80%
at the surface to 20- four-hundredth at three hundred hPa (9 km).
3.2 Modeling of Tropospheric Delays
The following sections shortly define the principles on that tropospheric delay modeling is predicated.
meteorologic inputs that ar necessary for the delay models, their vertical distribution and also the relations
between them ar treated in chapter four in addition.
It ought to be emphasised that the term tropospheric propagation delay is employed here despite the actual
fact that the hydrostatic part is additionally influenced by gases higher than the layer.
3.2.1 Generalized practical Description
The geometrical distance differs from the particular ray path by the distinction
n ( s ) .ds

SiA[NEU] =

ds

VAC

ATM

dSiA[NEU]: total/neutral slant path delay from receiver antenna A to satellite i

n: index of atmospherical refraction
ds: differential increment in distance with reference to the road of sight
ATM: ray path passing from antenna in direction to satellite through the atmosphere
VAC: virtual path of a ray passing from antenna in direction to satellite through vacuum.

SiA[NEU] = SiA[HYD] + SiA[WET]

S[HYD]: hydrostatic slant path delay
S[WET]: non-hydrostatic or wet slant path delay
a hydrostatic and a non-hydrostatic part. The latter is commonly referred to as wet part that isn't too wrong
because it is especially caused by the vertical distribution of water vapour within the layer and also the initial
part is additionally referred to as dry delay that is partially dishonest . a comparatively new side in
tropospheric delay modeling is to any distinguish between the azimuthally rhombohedral delay and uneven
components.

SiA[NEU] =SiA[HYD,SYMM] +SiA[HYD,SYMM] +SiA[WET,SYMM]+SiA[WET,SYMM]

S[SYMM]: tropospheric delay term under the assumption of symmetry in azimuth.
S[ASYMM]: tropospheric correction term taking asymmetric effects into account.
where the uneven parts ar typically determined by application of a horizontal tropospheric gradient model.
Consequently, the total notation for the neutral delay is
13

SiA[NEU] = m

i
A[HYD]

.ZHDA+m

.[GA[N,YD]. cos

i
A [AZI,HYD]

i
a

+GA[E,HYD].sin

].

m(): mapping function

ZHD: celestial point hydrostatic delay
ZWD: celestial point wet delay
G[N]: gradient in northward direction
G[E]: gradient in eastward direction
3.2.1 Modeling celestial point Delays
Following THAYER , the neutral celestial point path delay will be derived from the radio index of refraction
of the air
1

N =K .

Pd
T

-1
d

+ K2.

e
T

.Z

-1
w

+K3.

e
T2

-1
.Z w

n(s): index of refraction as operate of the space 's'

N: reduced index of tropospheric refraction
k1..3: refraction constants
Zd/w-1: inverse softness factors for dry and wet air
pd: dry pressure; metallic element = p - e with p being the entire pressure (measured quantity)
e: partial water vapour pressure

pi: pressure of perfect gas i

T: temperature
V: volume
m: mass
Mi: molar mass of gas i
Ri: specific universal gas constant
R0: universal gas constant; R0 = eight.31434 [J mol-1 K-1]

i: density of gas i

pi iRi T Zi
Zi: softness issue of gas i .
3.2.1.1 celestial point hydrostatic Delay

14

Three hydrostatic delay models are bestowed during this section. The Hopfield and also the Saastamoinen
model ar mentioned well and compared with one another. moreover, the MOPS model that doesn't want any
meteorologic measurements as input is delineated .
3.2.1.1.1Hopfield hydrostatic Delay Model
The equation for hydrostatic equilibrium follows from the best gas laws of Boyle- Mariotte and physicist
and might be expressed in differential type as

dp: differential amendment in pressure

g : gravity, assumed to be constant, esp. with height
: density of dry air, assumed to be constant
dH: differential amendment tall
and the density can even be expressed by

and ends up in the equation

P : pressure
T : temperature
Rd: specific universal gas constant of dry air.

As already mentioned, the vertical evolution of the temperature within the layer will be approximated by a
linear trend victimization the temperature lapse rate

T: temperature as operate of altitude 'f(h)'

T0: temperature at surface (or antenna) height h0 = zero m
: temperature lapse rate
H: height higher than water level
and is employed as model for the complete atmosphere. The equation becomes

15

k1: hydrostatic deflection constant (k1 = seventy seven.6 h/hPa)

R0: universal universal gas constant (R0 = eight.31434 J mol-1 K-1)
: total density
Md: molar weight of dry air (Md = twenty eight.9644 kg/kmol)
For a default temperature lapse rate of b = -6.81 K/km, a molar weight of dry air of Md = twenty eight.9644
kg/kmol, a universal universal gas constant of R0 = eight.31434 [Jmol-1K-1] and a mean gravity
acceleration of g = nine.806 m/s the exponent becomes h = four.02 four that's adopted for the Hopfield
two-quartic model. Another modification is formed by substitution of d

: temperature lapse rate in [K/km]

T0: surface/antenna temperature in [K]
t0: surface/antenna temperature in [C]
Hd: effective height of the dry atmosphere higher than the surface in [km]
Hd0: effective height of the dry atmosphere for a temperature of zero C in [km]
with the inverse negative height Hd that is that the effective height for the hydrostatic part and was obtained
by a work of worldwide radiosonde information to.

16

The celestial point hydrostatic delay of the Hopfield model will currently be written in closed type as

and can be computed with information of surface temperature and pressure3. As has become clear, variety of
approximations cause this easy formula. with the exception of the actual fact that the air is treated as a
perfect gas that isn't too essential, the idea of a continuing temperature lapse rate is one vital approximation
that ought to be stressed in addition because the incontrovertible fact that the gravity isn't sculptured with
dependence on the peak. consecutive model bestowed can overcome this disadvantage.
3.2.1.1.3 MOPS hydrostatic Delay Model
If no meteorologic information ar offered, the MOPS tropospheric formula will be applied. This approach
uses commonplace meteorologic information obsessed on latitude and takes differences due to the season
into consideration. First, the latitude-dependent mean meteorologic parts are taken from Table 3-2 and
afterwards denoted as x0. beginning with one January, the day of year DoY has got to be computed so as to
account for seasonal changes and also the corresponding values are to be taken from Table 3-3 and denoted
as Dx.

Latitude |
|

p0 [hPa]

T0 [K]

e0 [hPa]

0 [K/m]

0 [/]

1013.25

299.65

26.31

0.00630

2.77

1017.25

294.15

21.79

0.00605

3.15

30
45

1015.75

283.15

11.66

0.00558

2.57

60

1011.75

272.15

6.78

0.00539

1.813

75

1013.00

263.65

4.11

0.00453

1.55

<15

3.2.1.2 celestial point Wet Delay

17

Zenith wet delays ar full of the distribution of water vapour. it's terribly tough to derive extremely correct
models victimization solely surface measurements for this part, however it's potential to estimate celestial
point wet delays.

as extra parameters throughout network filtering. This issue are coated by section three.4. yet, some
approaches for the approximation of celestial point wet delays ar given here which will serve to predict wet
delays so as to initialize the Kalman filter.
3.2.2 projected celestial point Delays into Slant Direction
The tropospheric delay is shortest in celestial point direction and can become larger with increasing celestial
point angle. Projection of celestial point path delays into slant direction is performed by application of a
mapping operate or obliquity issue that's outlined as

S[NEU]: neutral/total delay in slant direction

ZND : zenith neutral delay
M
: mapping function
Z
: zenith angle from ground station to GPS satellite
for the neutral part, as an example, and is well found to be

SND: slant neutral delay, identical with S[NEU]

18

CHAPTER - 4
GPS Validation Experiments
Several hand-picked GPS validation experiments ar mentioned during this chapter. All tropospheric
parameters were calculable with the analysis system TropAC shaft that was developed as a part of this study.
As so much because the tropospheric delays ar involved, 3 levels of comparison ar potential.
The first level is that the comparison of celestial point neutral delays. Note that total delays aren't calculable
by the Trop AC analysis computer code, however the celestial point wet delays. However, the hydrostatic
delay is sculptured with facilitate of surface pressure and thus, it's not tough to derive the neutral delay.

ZND ZWDZHD
ZND : celestial point neutral (total) delay
ZHD : celestial point hydrostatic delay (modelled/predicted by pre-processor)
ZWD: celestial point wet delay (estimated by Kalman filter)
19

but mind the error budget: The variance for the neutral delay should be smaller than the variance of the wet
delay

D = WD + 2ZHD
because GPS alone is not able to sense wet delays. Instead, the hydrostatic component must be derived from
external information including additional measurement and modelling uncertainties. Valuable reference data
for this validation level come from the IGS analysis centers that provide a combined tropospheric product.
These combined data sets are total delays in zenith direction that are derived as the weighted mean of the
individual solutions of the particular analysis centers.
The second level is the comparison of zenith wet delays which are an output of the Trop AC software.
Reference data can be taken from radio sonde launches, for instance: The wet, vertical refractivity profile is
integrated and yields zenith wet delays.
Finally, the third level of comparison is the most interesting one for meteorology and climate research,
because it focuses on integrated water vapour (IWV) or precipitable water:

PW: precipitable water

Q: conversion factor to transform zenith wet delays into precipitable water and vice versa

The uncertainty of the conversion factor Q is an additional contributor to the total error budget of
precipitable water, but in most cases of minor concern. Water vapour radiometers can supply reference data
for water vapor comparisons.

Long-term Experiment
The long-term experiment was conducted for a period of approximately half a year. The basic characteristics
and objectives are given in Table 6-1 and a network plot with the nominal baseline configuration is given in
Figure 6-1. The primary objective of this experiment was to validate the tropospheric zenith delays that are
filtered in routine, automatic processing mode. For this reason, the standard configuration settings were
applied. Apart from accuracy considerations ,the availability of the observation data and the reliability of the
measurements are addressed in the following sections.

20

CHAPTER-5
SOPAC
The Scripps Orbit and Permanent Array Center (SOPAC) is located at the Cecil H. and Ida M. Green
Institute of Geophysics and Planetary Physics (IGPP), Scripps Institution of Oceanography (SIO), University
of California, San Diego (UCSD) in La Jolla, California. SOPAC's primary scientific role is to support high
precision geodetic and geophysical measurements using Global Positioning System (GPS) satellites,
particularly for the study of earthquake hazards, tectonic plate motion, plate boundary deformation, and
meteorological processes. SOPAC investigators also conduct research on the implementation, operation and
scientific applications of continuously monitoring GPS arrays and Synthetic Aperture Radar (SAR)
interferometry.
SOPAC is a major participant in the International GPS Service (IGS), serving as a Global Data Center and a
Global Analysis Center. SOPAC is the archive for the 250-station Southern California Integrated GPS
Network (SCIGN), as well as an analysis center and a data retrieval facility. SOPAC maintains an archive of
regional continuous GPS data for UNAVCO, Inc. focusing on the Western North America plate boundary
and is developing a seamless archive interface for UNAVCO. SOPAC performs research and provides
operational support for NOAA's Forecast Applications Branch (FAB) real-time GPS meteorology project for
21

short-term weather forecasting. SOPAC also provides essential infrastructure support for the California
Spatial Reference Center (CSRC), a partnership with the National Geodetic Survey (NGS).
SOPAC provides precise, rapid, ultra-rapid, and hourly orbits for the International GPS Service (IGS) and
NOAA's Forecast Systems Laboratory (FSL). Many GPS-related services and tools are available from the
SOPAC Web site, such as SCOUT (an ITRF Coordinates Generator; 15 minute turnaround time), Site
Information Manager (SIM), and a utility to check for unused 4-character site codes. SOPAC archives 24hour RINEX data from about 800 continuous GPS sites from more than 20 scientific networks around the
world, including the 250-station Southern California Integrated GPS Network (SCIGN). SOPAC also helps
maintain the operational center for the California Spatial Reference Center (CSRC). The CSRC provides
California's geodetic framework for scientific, surveying, engineering, and geographical information systems
in partnership with the National Geodetic Survey and Caltrans. SOPAC also collects and archives high-rate
(1 Hz), low latency (1-2 seconds) GPS data from stations in California.

CHAPTER -6
SUMMARY AND CONCLUSION
2.11
OBSERVATION DATA
G (GPS)
RINEX VERSION / TYPE
teqc 2015Jun23
gpsops
20151105 00:10:06UTCPGM / RUN BY / DATE
Linux 2.4.21-27.ELsmp|Opteron|gcc -static|Linux x86_64|=+
COMMENT
BIT 2 OF LLI FLAGS DATA COLLECTED UNDER A/S CONDITION
COMMENT
IISC
MARKER NAME
22306M002
MARKER NUMBER
GGN
C-MMACS
OBSERVER / AGENCY
ZR520013802
ASHTECH UZ-12
CQ00
REC # / TYPE / VERS
CR6200548032
ASH701945E_M
NONE
ANT # / TYPE
1337936.4550 6070317.1261 1427876.7852
APPROX POSITION XYZ
0.0780
0.0000
0.0000
ANTENNA: DELTA H/E/N
1
1
WAVELENGTH FACT L1/2
7
L1
L2
P1
P2
C1
S1
S2
# / TYPES OF OBSERV
30.0000
INTERVAL
COMMENT
This data is provided as a public service by NASA/JPL.
COMMENT
No warranty is expressed or implied regarding suitability
COMMENT

22

for use. For further information, contact:

Dave Stowers, NASA/JPL m/s 238-600
4800 Oak Grove Drive, Pasadena CA 91109 USA
Forced Modulo Decimation to 30 seconds
SNR is mapped to RINEX snr flag value [0-9]
L1 & L2: min(max(int(snr_dBHz/6), 0), 9)
pseudorange smoothing corrections not applied
2015
11
4
0
0
0.0000000
15 11 4 0 0
-20857289.06549
54.0004
-18848440.44649
54.0004
-11083435.89048
49.0004
-7765724.50947
46.0004
-7855683.72247
46.0004
-13101104.74148
52.0004
-5896475.03247
47.0004

GPS

COMMENT
COMMENT
COMMENT
COMMENT
COMMENT
COMMENT
COMMENT
COMMENT
TIME OF FIRST OBS
END OF HEADER

0.0000000 0 7G06G30G28G13G07G17G02
-16209735.50248 20797510.6104
20797515.1984
50.0004
-14666115.99748 20914110.0404
20914114.1794
50.0004
-8604092.59747 22277271.2704
22277273.7324
45.0004
-6042571.27546 23036761.2744
23036763.1584
41.0004
-6106433.23847 22983090.8734
22983092.0644
42.0004
-10148099.52848 21944905.1944
21944905.9324
48.0004
-4574999.56347 22902836.5494
22902836.6964
42.0004

20797510.033
20914110.113
22277271.943
23036761.759
22983091.305
21944904.986
22902837.669

1.0
COMPACT RINEX FORMAT
CRINEX VERS
/ TYPE
RNX2CRX ver.4.0.6
12-Nov-15 13:45
CRINEX PROG / DATE
2.11
OBSERVATION DATA
G (GPS)
RINEX VERSION / TYPE
teqc 2015Jun23
gpsops
20151105 00:10:06UTCPGM / RUN BY / DATE
Linux 2.4.21-27.ELsmp|Opteron|gcc -static|Linux x86_64|=+
COMMENT
BIT 2 OF LLI FLAGS DATA COLLECTED UNDER A/S CONDITION
COMMENT
IISC
MARKER NAME
22306M002
MARKER NUMBER
GGN
C-MMACS
OBSERVER / AGENCY
ZR520013802
ASHTECH UZ-12
CQ00
REC # / TYPE / VERS
CR6200548032
ASH701945E_M
NONE
ANT # / TYPE
1337936.4550 6070317.1261 1427876.7852
APPROX POSITION XYZ
0.0780
0.0000
0.0000
ANTENNA: DELTA H/E/N
1
1
WAVELENGTH FACT L1/2
7
L1
L2
P1
P2
C1
S1
S2
# / TYPES OF OBSERV
30.0000
INTERVAL
COMMENT
This data is provided as a public service by NASA/JPL.
COMMENT
No warranty is expressed or implied regarding suitability
COMMENT
for use. For further information, contact:
COMMENT
Dave Stowers, NASA/JPL m/s 238-600
COMMENT

23

4800 Oak Grove Drive, Pasadena CA 91109 USA

Forced Modulo Decimation to 30 seconds
SNR is mapped to RINEX snr flag value [0-9]
L1 & L2: min(max(int(snr_dBHz/6), 0), 9)
pseudorange smoothing corrections not applied
2015
11
4
0
0
0.0000000
&15 11

0.0000000

GPS

COMMENT
COMMENT
COMMENT
COMMENT
COMMENT
COMMENT
TIME OF FIRST OBS
END OF HEADER

7G06G30G28G13G07G17G02

3&-20857289065 3&-16209735502 3&20797510610 3&20797515198 3&20797510033 3&54000

3&50000 49484 4
4 4
3&-18848440446 3&-14666115997 3&20914110040 3&20914114179 3&20914110113 3&54000
3&50000 49484 4
4 4
3&-11083435890 3&-8604092597 3&22277271270 3&22277273732 3&22277271943 3&49000
3&45000 48474 4
4 4
3&-7765724509 3&-6042571275 3&23036761274 3&23036763158 3&23036761759 3&46000
3&41000 47464 4
4 4
3&-7855683722 3&-6106433238 3&22983090873 3&22983092064 3&22983091305 3&46000
3&42000 47474 4
4 4
3&-13101104741 3&-10148099528 3&21944905194 3&21944905932 3&21944904986 3&52000
3&48000 48484 4
4 4
3&-5896475032 3&-4574999563 3&22902836549 3&22902836696 3&22902837669 3&47000
3&42000 47474 4
4 4

CHAPTER -7
Bibliography
SCHUELER, T., G. W. HEIN, and B. EISSFELLER [2000b], Towards an Optimal Strategy for
GPS Wet Delay Kalman Filtering, Proceedings of IAIN 2000/ION Annual Meeting, San
Diego, Catamaran Hotel, 26-28 June 2000
SCHUELER, T., G. W. HEIN, and B. EISSFELLER [2000c], Improved Tropospheric Delay
Modeling Using an Integrated Approach of Numerical Weather Models and GPS,
Proceedings of ION GPS 2000, The Institute of Navigation, Salt Lake City, Utah, USA,
19-22 September 2000
SEEBER, G. [1989]; Satellitengeodsie, Verlag de Gruyter, Berlin - New York, 1993
SEEBER, G. [1993], Satellite Geodesy - Foundations, Methods, and Applications, Berlin;
New York: de Gruyter, 1993, ISBN 3-11-012753-9
24

 SELA, J. G. [1980], Spectral Modeling at the National Meteorological Center, Monthly

Weather Review, Vol. 108, pp. 1279-1292, September 1980
SOLBRIG, P. [2000], Untersuchungen ber die Nutzung numerischer Wettermodelle zur
Wasserdampfbestimmung mit Hilfe des Global Positioning Systems, Diploma Thesis,
Institute of Geodesy and Navigation, University FAF Munich, Germany, Feb. 2000
SPILKER, J. J. [1996], Tropospheric Effects on GPS, in: SPILKER and PARKINSON (eds.),
GPS Theory and Applications; Vol. I, Progress in Astronautics and Aeronautics, Vol.
163, pp. 517-546, 1996

25