Geodetic Science Reports (School of Earth Sciences)
http://hdl.handle.net/1811/24485
2017-01-16T22:07:52ZINS, GPS, and Photogrammetry Integration for Vector Gravimetry Estimation
http://hdl.handle.net/1811/78671
INS, GPS, and Photogrammetry Integration for Vector Gravimetry Estimation
Dwaik, Fathi Y., 1964-
Vector gravimetry using Inertial Navigation System (INS) in semi-kinematic
mode has been successfully applied. The integration of INS with other sensors, Global
Positioning System (GPS) or Gradiometer, for instance, has been under investigation for
many years. This dissertation examines the effect of photogrammetric derived orientation
on the INS sensor’s calibration and estimation of the gravity vector. The capability of
such integration in estimating the INS biases and drifts is studied. The underlying
principle, mathematical models, and error sources are presented and analyzed. The
estimation process utilizes the measurements of the Litton LN-100 inertial system,
Trimble 4000 SSI GPS dual frequency receiver, and metric frame camera. An optimal
filtering technique is used to integrate both GPS and INS on the level of raw
measurement for both systems. Introducing accurate and independent orientation
parameters, e.g., the photogrammetric source in this study, is demonstrated to enable
calibration of inertial gyros and bounding of their drift errors. This leads to improvement
in the horizontal components of the gravity vector estimation. The estimability and
improvement of the deflection of the vertical components are tested using flight test data
over Oakland, California, and a set of photogrammetric images simulated along the flight
trajectory.
The error statistics of the orientation measurement are modeled on the basis of the
variance-covariance matrix of a photogrammetric bundle adjustment of all photos. With
just a few ground control points at the beginning of the trajectory, the orientation
measurement errors along the trajectory are correlated significantly from epoch to epoch,
thus reducing the information content of the external orientation estimates.
The horizontal gravity component estimation is tested with respect to its
sensitivity to the variance of the orientation measurement errors, to its auto-correlation in
time, to the cross-correlation between angles, and to the amount of available ground
control. Although photogrammetric measurements, if uncorrelated, control orientation
errors as well as better than achievable with aircraft maneuvers, the inherent correlation
with a very limited amount of ground control provides only a small improvement. On the
basis of the simulation parameters, the gravity estimation error was reduced from 20
mgal (GPS/INS only) to about 9 mgal (best uncorrelated control) versus 17 mgal
(correlated control).
Presented in Partial Fulfillment of the Requirement for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University.; This work was supported by the U.S. Air Force under contract F19628-95-K- 0020 (Defense Mapping Agency funding) and by the National Imagery and Mapping Agency (formerly DMA) under contract NMA202-98-1-1110.
1998-01-01T00:00:00ZDwaik, Fathi Y., 1964-Automatic Recognition and Location of Civil Infrastructure Objects Using Mobile Mapping Technology, Neural Network and Markov Chain Monte Carlo
http://hdl.handle.net/1811/78670
Automatic Recognition and Location of Civil Infrastructure Objects Using Mobile Mapping Technology, Neural Network and Markov Chain Monte Carlo
Li, Ron; Tu, Zhuowen
Project Report (November 1998 – December 1999). Submitted to The OSU Center For Mapping.
2000-01-01T00:00:00ZLi, RonTu, ZhuowenGRACE Time-Variable Gravity Field Recovery Using an Improved Energy Balance Formalism
http://hdl.handle.net/1811/78669
GRACE Time-Variable Gravity Field Recovery Using an Improved Energy Balance Formalism
Shang, Kun
Earth’s gravity is continuously varying with respect to time due primarily to mass
transports within the Earth system and external gravitational forcing. A new formalism
based on energy conservation principle for time-variable gravity field recovery using
satellite gravimetry has been developed and yields more accurate estimation of in-situ
geopotential difference observables using K-Band Ranging (KBR) measurements from
the Gravity Recovery and Climate Experiment (GRACE) twin-satellite mission. The new
approach can preserve more time-variable gravity information sensed by KBR range-rate
measurements and reduce orbit error as compared to previous energy balance studies.
Results based on analysis of more than 10 years of GRACE data indicate that the
estimated geopotential differences agree well with the predicted values from official
Level 2 solutions: with much higher correlation of 0.9, as compared to 0.5–0.8 reported
by previous energy balance studies. This study demonstrates that the new approach is
more flexible for both global and regional temporal gravity recovery, leading to the first
independent GRACE monthly solution series based on energy conservation principle,
which is comparable to the results from different approach. The developed formalism is
applicable to the general case of low-low satellite-to-satellite radiometric or laser
interferometric tracking measurements, such as GRACE Follow-on or other Next
Generation Gravity Field missions, for efficient retrieval and studies of Earth’s mass
transport evolutions.
The regional gravity analysis over Greenland reveals that a substantially higher temporal
resolution is achievable at 10 or 11-day interval from GRACE data, as compared to the
official monthly solutions, but without the compromise of spatial resolution, nor the need
to use regularization or post-processing. Studies of the terrestrial and ground water
storage change over North China Plain show high correlation in sub-monthly scale,
among the 11-day time-variable gravity solutions from this study, in-situ data, and
hydrologic and atmospheric models. The 11-day solutions with 1-day step successfully
capture the surface mass change caused by the rapid snow and ice accumulation and
melting during the extreme weather event of 2008 Southeast China snow and ice storm.
These results demonstrated that sub-monthly solutions from GRACE can provide an
additional constraint to understand the rapid mass transport and the dynamic processes
for both extreme weather events and short-time surface and ground water monitoring,
which may potentially improve our understanding of various mass transports within the
Earth system, and applicable to societal services such as disaster response or mitigation,
and water resources management.
This Report was prepared for and submitted to the Graduate School of the Ohio State University as a dissertation in partial fulfillment of the requirements for the PhD degree.; This research is conducted under the supervision of Professor C.K. Shum, Division of Geodetic Science, School of Earth Sciences, The Ohio State University. This research is partially supported by grants from NSF via the Belmont Forum/IGFA Grant (ICER- 1342644), and NASA’s geodesy and cryosphere grants (NNX12AJ95G, NNX12AK28G, NNX11AR47G). GRACE data products are from NASA’s PODAAC via Jet Propulsion Laboratory/California Institute of Technology (JPL), University of Texas Center for Space Research (CSR), and GeoForschungsZentrum Potsdam (GFZ). Some figures in this paper were generated using the Generic Mapping Tools (GMT) [Wessel and Smith, 1991]. The computational aspect of this work was supported in part by an allocation of computing resources from the Ohio Supercomputer Center (http://www.osc.edu).
2015-08-01T00:00:00ZShang, KunSource Parameters Inversion for Recent Large Undersea Earthquakes from GRACE Data
http://hdl.handle.net/1811/78668
Source Parameters Inversion for Recent Large Undersea Earthquakes from GRACE Data
Dai, Chunli
The north component of gravity and gravity gradient changes from the Gravity Recovery
And Climate Experiment (GRACE) are used to study the coseismic gravity change for
five earthquakes over the last decade: the 2004 Sumatra-Andaman earthquake, the 2007
Bengkulu earthquake, the 2010 Maule, Chile earthquake, the 2011 Tohoku earthquake,
and the 2012 Indian Ocean earthquakes. We demonstrate the advantage of these north
components to reduce north-south stripes and preserve higher spatial resolution signal in
GRACE Level 2 (L2) monthly Stokes Coefficients data products. By using the high
spherical harmonic degree (up to degree 96) data products and the innovative GRACE
data processing approach developed in this study, the retrieved gravity change is up to –
34±1.4 μGal for the 2004 Sumatra and 2005 Nias earthquakes, which is by far the highest
coseismic signal retrieved among published studies. Our study reveals the detectability of
earthquakes as small as Mw 8.5 (i.e., the 2007 Bengkulu earthquake) from GRACE data.
The localized spectral analysis is applied as an efficient method to determine the practical
spherical harmonic truncation degree leading to acceptable signal-to-noise ratio, and to
evaluate the noise level for each component of gravity and gravity gradient change of the
seismic deformations.
By establishing the linear algorithm of gravity and gravity gradient change with respect
to the double-couple moment tensor, the point source parameters are estimated through
the least squares adjustment combined with the simulated annealing algorithm. The
GRACE-inverted source parameters generally agree well with the slip models estimated
using other data sets, including seismic, GPS, or combined data. For the 2004 Sumatra-
Andaman and 2005 Nias earthquakes, GRACE data produce a shallower centroid depth
(9.1 km) compared to the depth (28.3 km) from GPS data, which may be explained by the
closer-to-trench centroid location and by the aseismic slip over the shallow region. For
the 2011 Tohoku earthquake, the inversions from two different GRACE data products
and two different forward modeling produce similar source characteristics, with the
centroid location southwest of and the slip azimuth 10° larger than the GPS/seismic
solutions. The GRACE-estimated dip angles are larger than that from GPS/seismic data
for the 2004 Sumatra-Andaman and 2005 Nias earthquakes, the 2010 Maule, Chile
earthquake, and the 2007 Bengkulu earthquake. These differences potentially show the
additional offshore constraint from GRACE data, compared to GPS/seismic data. With
more accurate and higher spatial resolution measurements anticipated from the GRACE
Follow-on mission, with a scheduled launch date in 2017, we anticipate the data will be
sensitive to even smaller earthquake signals. Therefore, GRACE type observations will
hopefully become a more viable measurement to further constrain earthquake focal
mechanisms.
This Report was prepared for and submitted to the Graduate School of the Ohio State University as a dissertation in partial fulfillment of the requirements for the PhD degree.; This research is conducted under the supervision of Professor C.K. Shum, Division of Geodetic Science, School of Earth Sciences, The Ohio State University. This research is primarily supported by NASA’s Earth and Space Science Fellowship (ESSF) Program (Grant NNX12AO06H), partially supported by National Science Foundation (NSF) Division of Earth Sciences (Grant EAR-1013333). GRACE data products are from NASA’s PODAAC via Jet Propulsion Laboratory/California Institute of Technology (JPL), University of Texas Center for Space Research (CSR), and GeoForschungsZentrum Potsdam (GFZ). Preliminary GPS time series provided by the ARIA team at JPL and Caltech. All original GEONET RINEX data were provided to California Institute of Technology by the Geospatial Information Authority (GSI) of Japan. Some figures in this paper were generated using the Generic Mapping Tools (GMT) [Wessel and Smith, 1991]. This work was also supported in part by an allocation of computing resources from the Ohio Supercomputer Center (http://www.osc.edu).
2015-08-01T00:00:00ZDai, ChunliHeights, the Geopotential, and Vertical Datums
http://hdl.handle.net/1811/78667
Heights, the Geopotential, and Vertical Datums
Jekeli, Christopher
This report reviews the fundamental definitions of heights and vertical datums, specifically
motivated by the modern technique of determining heights using accurate satellite vertical
positioning in combination with an accurate model for the geopotential. It is shown that the
determination of heights in such a manner requires knowledge of the potential value of the vertical
datum (as opposed to leveling procedures that do not require this). Furthermore, to determine the
potential of a vertical datum ideally requires normal heights (defined at the origin point of the
datum, or determined elsewhere by leveling) rather than orthometric heights, as this avoids the
complication of assuming a density model for the crust. The models associated with these
procedures are also developed within the context of the temporally varying field of the tidal
potential, which leads to a more fundamental distinction between a vertical datum (local geoid) and
the global geoid. That is, the global geoid, by definition, has always the same potential value but
its surface varies (varying geoid undulation); while the vertical datum has an origin changing only
because of crustal deformation and the potential varies due to the direct and indirect tidal effects.
The models thus developed also form the basis for monitoring the stability of vertical datums
under the influence of geodynamic vertical crustal deformations, such as caused by post-glacial
rebound. This has obvious implications in the monitoring of lake levels that are tied to a particular
vertical datum. Preliminary models and procedures are indicated.
Technical Report. Ohio Sea Grant Development Program, NOAA. Grant No. NA86RG0053 (R/CE-7-PD). "Modern Vertical Datums and Lake Level".
2000-11-01T00:00:00ZJekeli, ChristopherApplications of Parameter Estimation and Hypothesis Testing to GPS Network Adjustments
http://hdl.handle.net/1811/78666
Applications of Parameter Estimation and Hypothesis Testing to GPS Network Adjustments
Snow, Kyle Brian
It is common in geodetic and surveying network adjustments to treat the rank deficient normal equations in a way that produces zero variances for the so–called "control" points.
This is often done by placing constraints on a minimum number of the unknown parameters, typically by assigning a zero variance to the a priori values of these parameters (coordinates). This approach may require the geodetic engineer or analyst to make an arbitrary decision about which parameters to constrain, which may have undesirable effects, such as parameter error ellipses that grow with distance from the constrained point.
Constraining parameters to a priori values is only one way of overcoming the rank deficiency inherent in geodetic and surveying networks. There are more preferable ways, which this thesis presents, namely Minimum Norm Least–Squares Solution (MINOLESS) and Best Linear Minimum Partial Bias Estimation (BLIMPBE). MINOLESS not only minimizes the weighted norm of the observation error vector but also minimizes the norm of the parameter vector, while BLIMPBE minimizes the bias for a subset of the parameters. In this thesis, these techniques are applied to a geodetic network that serves as a datum access for GPS–buoy work in Lake Michigan. The GPS–buoy has been used extensively in recent years by NOAA, The Ohio State University (OSU), and other organizations to determine lake and ocean surface heights for marine navigation and scientific studies. The work presented in this paper includes 1) parameter estimation using (Weighted) MINOLESS and hypothesis testing for the purpose of determining if recent observations are consistent with published coordinates at an earlier epoch; 2) a discussion of the BLIMPBE estimation technique for three new points to be used as GPS–buoy fiducial stations and a comparison of this technique to the "Adjustment with Stochastic Constraints" method; 3) usage of standardized reliability numbers for correlated observations; 4) a proposal for outlier detection and minimum outlier computation at the GPS–baseline level. The work may also be used as an example to follow for establishing new fiducial points with respect to a geodetic reference frame using observed GPS baseline vectors.
The results of this work lead to the following conclusions: 1) MINOLESS is the parameter estimation techniques of choice when it is required that changes to all a priori coordinates be minimized while performing a minimally constrained adjustment; 2) BLIMPBE appears to be an attractive alternative for selecting subsets of the parameter vector to adjust. BLIMPBE solutions using various selection–matrix types are worthy of further investigation; 3) outlier detection at the GPS–baseline level permits the entire observed baseline to be evaluated at once, rather than making decisions regarding the ii
hypothesis at the baseline–component level. It is shown that the two approaches can yield different results.
This report was prepared by Kyle Snow while a student in the Department of Civil and Environmental Engineering and Geodetic Science. It was submitted to the Graduate School of The Ohio State University in the Autumn of 2002 in partial fulfillment of the requirements of the Master of Science degree. Prof. Burkhard Schaffrin served as advisor and Prof. C.K. Shum as co–advisor, both in the program for Geodetic Science and Surveying.; The research was funded in part by the Office of Naval Research Naval Oceanographic Partnership Program (NOPP), under the Ohio State University component of the Gulf of Mexico Monitoring System, and the NASA Physical Oceanography program under the TOPEX/POSEIDON Extended Mission project.
2002-12-01T00:00:00ZSnow, Kyle BrianLinear Features in Photogrammetry
http://hdl.handle.net/1811/78665
Linear Features in Photogrammetry
Habib, Ayman; Asmamaw, Andinet; Kelley, Devin; May, Manja
This research addresses the task of including points as well as linear features in
photogrammetric applications. Straight lines in object space can be utilized to perform
aerial triangulation. Irregular linear features (natural lines) in object space can be utilized
to perform single photo resection and automatic relative orientation.
When working with primitives, it is important to develop appropriate representations
in image and object space. These representations must accommodate for the perspective
projection relating the two spaces. There are various options for representing linear
features in the above applications. These options have been explored, and an optimal
representation has been chosen.
An aerial triangulation technique that utilizes points and straight lines for frame and
linear array scanners has been implemented. For this task, the MSAT (Multi Sensor
Aerial Triangulation) software, developed at the Ohio State University, has been
extended to handle straight lines. The MSAT software accommodates for frame and
linear array scanners.
In this research, natural lines were utilized to perform single photo resection and
automatic relative orientation. In single photo resection, the problem is approached with
no knowledge of the correspondence of natural lines between image space and object
space. In automatic relative orientation, the problem is approached without knowledge of
conjugate linear features in the overlap of the stereopair. The matching problem and the
appropriate parameters are determined by use of the modified generalized Hough
transform. These techniques were tested using simulated and real data sets for frame
imagery.
2000-01-01T00:00:00ZHabib, AymanAsmamaw, AndinetKelley, DevinMay, ManjaAirborne Vector Gravimetry Using GPS/INS
http://hdl.handle.net/1811/78664
Airborne Vector Gravimetry Using GPS/INS
Kwon, Jay Hyoun, 1967-
Compared to the conventional ground measurement of gravity, airborne
gravimetry is relatively efficient and cost-effective. Especially, the combination of GPS
and INS is known to show very good performances in the range of medium frequencies
(1-100 km) for recovering the gravity signal.
Conventionally, gravity estimation using GPS/INS was analyzed through the
estimation of INS system errors using GPS position and velocity updates. In this case,
the complex navigation equations must be integrated to obtain the INS position, and
the gravity field must be stochastically modeled as a part of the state vector. The
vertical component of the gravity vector is not estimable in this case because of the
instability of the vertical channel in the solution of the inertial navigation equations.
In this study, a new algorithm using acceleration updates instead of
position/velocity updates has been developed. Because we are seeking the gravitational
field, that is, accelerations, the new approach is conceptually simpler and more
straightforward. In addition, it is computationally less expensive since the navigation
equations do not have to be integrated. It is more objective, since the gravity
disturbance field does not have to be explicitly modeled as state parameters.
An application to real test flight data as well as an intensive simulation study has
been performed to test the validity of the new algorithm. The results from the real
flight data show very good accuracy in determining the down component, with
accuracy better than ±5 mGal. Also, a comparable result was obtained for the
horizontal components with accuracy of ±6 to ±8 mGal. The resolution of the final
result is about 10 km due to the attenuation with altitude.
The inclusion of a parametric gravity model into the new algorithm is also
investigated for theoretical reasons. The gravity estimates from this filter showed
strong dependencies on the model and required extensive computation with no
improvement over the approach without parametric gravity model.
This report was prepared by Jay Hyoun Kwon, a graduate student, Department of Civil and Environmental Engineering and Geodetic Science, under the supervision of Professor Christopher Jekeli.; This research was supported by the National Imagery and Mapping Agency (NIMA); Contract No. NMA202-98-1-1110.; It was submitted to the Graduate School of The Ohio State University in the Winter of 2000 in partial fulfillment of the requirements of the Doctor of Philosophy degree.
2000-04-01T00:00:00ZKwon, Jay Hyoun, 1967-Geophysical Investigations on Gravity Gradiometry and Magnetic Data over the Wichita Uplift Region, Southwestern Oklahoma
http://hdl.handle.net/1811/78663
Geophysical Investigations on Gravity Gradiometry and Magnetic Data over the Wichita Uplift Region, Southwestern Oklahoma
Erkan, Kamil
The Wichita uplift in southwestern Oklahoma is a unique region that shows strong gravity and
magnetic field anomalies. Detailed geologic data as well as structural cross sections are also
available for the region. This report includes a qualitative geophysical analysis of the airborne
gravity gradiometer profiles, and a quantitative analysis of an airborne magnetic field data
collected in the region. Two datasets were analyzed independently. Firstly, an effort has been
made by comparative analyses of different gravity gradient components with the gravity field
from EGM2008 side by side in order to understand the nature of the subsurface structural setting.
Secondly, a spectral analysis of magnetic field has been applied using the well-known power-law
behavior of the magnetic field. The resulting source intensity map delineates the areas with high
magnetic sources, and also is in agreement with the geologic findings in the region.
This report was prepared by Dr. Kamil Erkan for work completed while on a post-doctoral fellowship at the School of Earth Sciences, Ohio State University, with support from research funds in the Geodetic Science Program.
2015-04-01T00:00:00ZErkan, KamilGravity Recovery Using COSMIC GPS Data: Application of Orbital Perturbation Theory
http://hdl.handle.net/1811/78662
Gravity Recovery Using COSMIC GPS Data: Application of Orbital Perturbation Theory
Hwang, Cheinway
COSMIC is a joint Taiwan-US mission to study atmosphere using GPS occultation. Its GPS data
for precise orbit determination can be used for gravity recovery. In this report a kinematic
approach was employed which assumes the positional data can be derived from the GPS data of
COSMIC in the operational phase. Using the geometric relationship between the positional
variations of orbit and the variations in the six Keplerian elements, improved formulae for the
radial, along-track and cross-track perturbations were derived. Based on a comparison with true
perturbations from numerical integrations, these formulae are more accurate than the commonly
used order-zero formulae. The improved formulae were used to simulate gravity recovery using
the COSMIC data. In one simulation with the OSU91A model to degree 50 as the a priori
geopotential model, it is demonstrated that the EGM96 model can be improved up to degree 26
using one year of COSMIC data.
A significant effort was devoted to the recovery of temporal gravity variation using
COSMIC data. Sea level anomaly (SLA) was first generated using the Cycle 196
TOPEX/POSEIDON altimeter data. The steric anomaly due to thermal expansion was created
using temperature data at 14 oceanic layers. The steric anomaly-corrected SLA was used to
generate harmonic coefficients of temporal gravity variation. With a 3-cm noise at a one-minute
sampling interval in the COSMIC data, the gravity variation cannot be perfectly reproduced, but
the recovered field clearly shows the gravity signature due to mass movement in an El Niño.
With a 0.1-cm noise, the temporal gravity variation up to harmonic degree 10 is almost exactly
recovered and this prompts the need of a better processing technique and a sophisticated GPS
receiver technology.
The idea described in this report was initiated during the author's visit to the Ohio State University in the
Summer of 1998, hosted by Prof. C.K. Shum.; This research was partly supported by the National Science Council of ROC.
1998-10-01T00:00:00ZHwang, CheinwayStatic and Kinematic Absolute GPS Positioning and Satellite Clock Error Estimation
http://hdl.handle.net/1811/78661
Static and Kinematic Absolute GPS Positioning and Satellite Clock Error Estimation
Han, Shin-Chan, 1975-
This study presents the results of investigations to determine accurate position
coordinates using the Global Positioning System in the absolute (point) positioning mode.
The most common method to obtain accurate positions with GPS is to apply doubledifferencing
procedures whereby GPS satellite signals are differenced at a station and
these differences are again differenced with analogous differences at other stations. The
differencing between satellites eliminates the receiver clock errors, while the betweenstation
differences eliminate the satellite clock errors (as well as other errors, such as orbit
error). However, only coordinate differences can be determined in this way and the
accuracy depends on the baseline length between cooperating stations. The strategy with
accurate point positioning is to estimate GPS satellite clock errors independently, thus
obviating the between-station differencing. The clock error estimates are then used in an
application of a single-difference (between-satellite) positioning algorithm at any site to
determine the coordinates without reference to any other site. Using IGS (International
GPS Service) orbits and station coordinates, the GPS clock errors were estimated at 30-
second intervals and these estimates were compared to values determined by JPL
(Zumberge et al., 1998). The agreement was at the level of about 0.1 nsec (3 cm).
The absolute positioning technique was tested in an application of a single-differenced
(between-satellite) positioning algorithm in static and kinematic modes. For the static
case, an IGS station was selected and the coordinates were estimated. The estimated
absolute position coordinates and the published values had a mean difference of up to 18
cm with standard deviation less than 2 cm. For the kinematic case, data (every second)
obtained from a GPS buoy were tested and the result from the absolute positioning was
compared to a DGPS solution. The mean difference between the two algorithms is less
than 40 cm and the standard deviation is less than 23 cm. It was proved that a higher rate
(less than 30 sec.) of satellite clock determination and a good tropospheric delay model
are required to do absolute kinematic positioning to better than 10 cm accuracy.
This report was prepared by Shin-Chan Han, a graduate student, Department Civil and Environmental Engineering and Geodetic Science, under the supervision of Professor Christopher Jekeli. This research was supported by the National Imagery and Mapping Agency under Air Force Phillips Laboratory contracts F19628-95-K-0020 and F19628- 96-C-0169.; This report was also submitted to the Graduate School of Ohio State University as a thesis in partial fulfillment of the requirements for the Master of Science degree.
2000-04-01T00:00:00ZHan, Shin-Chan, 1975-Static Calibration of Tactical Grade Inertial Measurement Units
http://hdl.handle.net/1811/78660
Static Calibration of Tactical Grade Inertial Measurement Units
Hayal, Adem G.
The demand for precise positioning grows up parallel to the advances in production of the
geolocation instruments. Today, the Global Positioning System (GPS) is the most
common positioning system in use because of its being very precise, convenient and
cheap. However, when working in such areas that the external references (e.g. GPS
satellites) are not available, a system that does not require information from any external
source of information is required. Especially, these kinds of systems necessitate in
detection of unexploded ordnances (UXO) buried in forestry areas, where precise
position information is vital for removing them. The Inertial Navigation System (INS)
operates in any environment and does not depended on any external source of
information. It can operate alone or as an integrated system with GPS. However, the
Inertial Measurement Unit (IMU) sensor outputs include some errors which can cause
very large positioning errors. These errors can significantly be reduced by using
calibration methods. The most accurate calibration methods are performed in laboratories
and they require very precise instruments. However, the most significant IMU errors,
biases and scale factor errors, change from turn on to turn on of the IMU and therefore
they need to be estimated before every mission. The Multi-Position Calibration Method
developed by Shin (2002) is a good example which is cost efficient and it can be applied
in the field without use of any external calibration instrument. The method requires
numerous IMU attitude measurements and use the gravity magnitude and Earth rotation
rate as reference for calibration.
The performance of the Multi-Position Calibration Method was tested by using a cart
based geolocation system which includes 2 tactical grade IMUs, Honeywell HG1700 and
HG1900. The calibration test was conducted in a parking lot of Ohio State University on
06 June 2010. The calibration estimations have shown that the navigation accuracy could
be improved by up to 19.8% for the HG1700 and 17.8% for the HG1900. However, the
results were not consistent among each other and in some cases decrease in the
positioning accuracy was yielded.
The research that led to this report was partially supported by a contract from the Strategic Environmental Research and Development Program (SERDP), contract number, W912HQ-08-C- 0044.; This report was also presented in partial fulfillment of the requirements for the Master of Science Degree to the Graduate School of the Ohio State University.
2010-09-01T00:00:00ZHayal, Adem G.Direct Sensor Orientation in Airborne and Land-based Mapping Applications
http://hdl.handle.net/1811/78659
Direct Sensor Orientation in Airborne and Land-based Mapping Applications
Grejner-Brzezinska, Dorota A.
Since the early 1990s, the concept of Mobile Mapping Systems (MMS) has evolved from rather
simple land-based systems to more sophisticated, real-time multi-tasking and multi-sensor
systems, operational in land and airborne environments. Mobile Mapping technology has made a
remarkable progress, notably expanding its use in remote sensing, and surveying and mapping
markets. New systems are being developed and built for specialized applications, in support of
land-based and airborne imaging sensors, aimed at automatic data acquisition for GIS databases.
The major objective of this report is to review the concept of Mobile Mapping System and
GPS/INS supported direct platform orientation (DPO) in particular, as well as their evolution
since early 1990s, with a special emphasis on the research and development carried out in this
area at the Ohio State University. A short review of the inertial navigation concept is given, and a
notion of GPS/INS (inertial navigation system) integration is also presented. The concept of
direct georeferencing is also explained and compared to the traditional aerotriangulation (AT)
method of image geo-registration, and the importance of multi-sensor system calibration is
discussed, including its impact on the positioning accuracy. Some examples of currently
attainable navigation performance, based on the OSU-developed Airborne Integrated Mapping
System (AIMS) and the land-based system for highway mapping are discussed, and future
perspectives of MMS are presented.
Although MMS may be, in general, associated with land-based applications, the concept of
airborne mapping (remote sensing) based on DPO is also discussed here, primarily due to the fact
that airborne positioning and orientation systems based on GPS/INS integration are based on
similar hardware and software designs, and clearly evolved from the traditional airborne mapping
as a consequence of the advent of a high-accuracy GPS/INS systems. Thus, the DPO facilitated
through GPS/INS fusion is a common denominator for the modern land-based and airborne
mapping, which often times involves multiple imaging sensors to achieve higher accuracy, and
data complementarity and redundancy.
The work described in this report is a summary of research and development in the area of
GPS/INS (Global Positioning System/Inertial Navigation System) integration for direct
georeferencing of imaging sensors to support precision airborne and land-based mapping. The
research was conducted at the Ohio State University Center for Mapping and the Department of
Civil and Environmental Engineering and Geodetic Science (CEEGS).; The work presented here was supported by the National Aeronautics and Space Administration
(NASA) grant, OSU # 733223, the Ohio Department of Transportation (ODOT) grant, OSU #
733488, and the Federal Department of Transportation (FDOT) grant, OSU # 739152.
2001-06-01T00:00:00ZGrejner-Brzezinska, Dorota A.Object Recognition from AIMS Data Using Neural Networks
http://hdl.handle.net/1811/78658
Object Recognition from AIMS Data Using Neural Networks
Li, Ron; Ma, Fei; Tu, Zhuowen
Project Report (November 1997 – December 1998). Submitted to The OSU Center For Mapping.
1998-12-01T00:00:00ZLi, RonMa, FeiTu, ZhuowenCoastal Altimetry and Applications
http://hdl.handle.net/1811/78657
Coastal Altimetry and Applications
Anzenhofer, Michael; Shum, C. K.; Rentsh, Mathias
This report was prepared by Dr. Michael Anzenhofer of the Geo-Forschungs-Zentrum (GFZ) Potsdam, Germany, while visiting the Department of Civil and Environmental Engineering and Geodetic Science (CEEGS), Ohio State University, during 1997-1998. The visit was hosted by Prof. C.K. Shum of the Department of Civil and Environmental Engineering and Geodetic Science.; This work was partially supported by NASA Grant No.735366, Improved Ocean Radar Altimeter and Scatterometer Data Products for Global Change Studies and Coastal Application, and by a grant from GFZ, Prof. Christoph Reigber, Director.
1999-01-01T00:00:00ZAnzenhofer, MichaelShum, C. K.Rentsh, MathiasSpline Representations of Functions on a Sphere for Geopotential Modeling
http://hdl.handle.net/1811/78653
Spline Representations of Functions on a Sphere for Geopotential Modeling
Jekeli, Christopher
Three types of spherical splines are presented as developed in the recent literature on
constructive approximation, with a particular view towards global (and local)
geopotential modeling. These are the tensor-product splines formed from polynomial and
trigonometric B-splines, the spherical splines constructed from radial basis functions, and
the spherical splines based on homogeneous Bernstein-Bézier (BB) polynomials. The
spline representation, in general, may be considered as a suitable alternative to the usual
spherical harmonic model, where the essential benefit is the local support of the spline
basis functions, as opposed to the global support of the spherical harmonics. Within this
group of splines each has distinguishing characteristics that affect their utility for
modeling the Earth’s gravitational field. Tensor-product splines are most
straightforwardly constructed, but require data on a grid of latitude and longitude
coordinate lines. The radial-basis splines resemble the collocation solution in physical
geodesy and are most easily extended to three-dimensional space according to potential
theory. The BB polynomial splines apply more generally to any sphere-like surface (e.g.,
the geoid or the Earth’s surface) and have a strong theoretical legacy in the field of spline
approximations. This report provides a review of these three types of splines, their
application to the geodetic boundary-value problem, and formal expressions for
determining the model coefficients using data with observational errors.
This report was prepared with support from the National Geospatial-Intelligence Agency
under contract NMA302-02-C-0002 and serves as the final technical report for this
project.
2005-03-01T00:00:00ZJekeli, ChristopherGPS Buoy Campaigns for Vertical Datum Improvement and Radar Altimeter Calibration
http://hdl.handle.net/1811/78652
GPS Buoy Campaigns for Vertical Datum Improvement and Radar Altimeter Calibration
Cheng, Kai-chien
This report summarized three Global Positioning System (GPS) buoy campaigns in the
Great Lakes from 1999 to 2003 that were carried out by the Laboratory of Space Geodesy and
Remote Sensing Research in the Department of Civil and Environmental Engineering and
Geodetic Science (CEEGS), at the Ohio State University. The report focuses on the field
work procedure of GPS buoy operation in these past campaigns and is intended to provide
experience for similar applications in the future.
The campaigns in this report include the Holland Campaign in Lake Michigan in 1999,
the Marblehead Campaign in Lake Erie in 2001, and the Cleveland Campaign in Lake Erie in
2003. The major objective of these campaigns is to establish a calibration site for multiple
satellite altimeters by using the GPS buoy to and the existing tide gauges provided by the
Center for Operational Oceanographic Products and Services (CO-OPS) in the National
Oceanic and Atmospheric Administration (NOAA). The campaigns provide useful information
to the applications including radar altimeter absolute calibration, the establishment of the safe
navigation in the Great Lakes, and the development of an integrated shoreline information in a
spatial information database for coastal management and decision making. Since the report
focuses primarily on the field work procedure, only limited results are presented. The published
calibration results using the data from these campaigns are cited in this report.
Generally, the GPS buoy is defined by putting GPS equipments on a floating object,
which includes different types of buoys and could even be a moving vessel. The use of GPS
buoys is a relatively new technique for the marine applications and its designs and operations
vary from one application to another. For example, its platform could range from a small lifesaver
type to an autonomous ruggedized type buoy. However only the OSU waverider GPS
buoy, a life-saver type buoy that was used in these campaigns, is stressed in this report.
The OSU waverider GPS buoy is a fairly simple design: it is built by attaching a
Dorne/Margolin Element with Choke Ring antenna on top of a 2- feet (diameter) life-saver buoy
covered with a transparent radome. The buoy is tethered to a boat where the receiver, power
supply and the operators reside. Marks are made on four sides of buoy and their offsets to the
antenna reference point (ARP) are carefully measured in the laboratory. The operator needs
to observe the water surface with respect to these marks in order to accurately refer ARP to the
water surface. The buoy data is post-processed with differential GPS (DGPS) in kinematic
mode after the field work.
The campaign-related documents, including National Geodetic Survey (NGS) data
sheets, GPS Station Observation Log, Visibility Obstruction Diagram, campaign proposal, and
field work log, are attached in the Appendices.
The report was prepared by Kai-chien Cheng, a graduate research associate in the
Laboratory for Space Geodesy and Remote Sensing, at the Ohio State University, under the
supervision of Professor C. K. Shum. This report was supported by the National Oceanographic
Partnership Program Grant (Dynalysis of Princeton #865618), National Aeronautics and Space
Administration TOPEX/POSEIDON Extended Mission Grant (NAG 5-6910/JPL961462),
National Aeronautics and Space Administration Earth Science Information Partnership CAN
Grant (CIT #12024478), National Aeronautics and Space Administration Interdisplanary
Science Project (NAG5-9335), National Science Foundation Digital Government Grant (EIA-
0091494, and the Ohio Sea Grant Program (R/CE-5).
2004-01-01T00:00:00ZCheng, Kai-chienStatistical Analysis of Moving-Base Gravimetry and Gravity Gradiometry
http://hdl.handle.net/1811/78651
Statistical Analysis of Moving-Base Gravimetry and Gravity Gradiometry
Jekeli, Christopher
Moving-base gravimetry systems require multiple sensors to extract the gravitational signal – an
accelerometer (or gravimeter), or a set of mutually orthogonal accelerometers that sense the action
forces on the vehicle; a suite of gyroscopes (or a stabilized platform) that provides proper
orientation for the accelerometers; and a geometric (kinematic) positioning system (e.g., GPS)
from which the kinematic acceleration may be derived, and that also provides geospatial
referencing of the signals. The error in the recovered gravitational signal depends on the individual
sensor errors, but also on the coupling of the sensor errors to the actual acceleration environment
of the system. The error analysis is fairly well known and documented in the literature and agrees
largely with experimental and operational results. This report reviews the analysis in detail and
extends it to moving-base gravity gradiometry. In the latter case the system comprises a set of
gradiometers (or differential accelerometers), a suite of gyros for orientation (stabilization), and a
geospatial referencing system (GPS). The errors in the recovered gravitational gradients depend
on the sensor errors, but also on the coupling of these errors to the angular rate environment of the
system. The analyses specifically target airborne systems used for gravity and gravity gradient
mapping. While the orientation bias error is especially detrimental to airborne gravimetry, it is the
random noise in the gyro angular rate that contributes most to airborne gradiometry, as it couples
with the total angular rate. The analysis shows that a gradiometer with 1 E/ √Hz sensitivity will
not be adversely compromised (at medium and high frequencies) if the required gyros have bias
repeatability of 0.0015 °/hr and sensitivity of 0.01 °/hr/ √Hz ≈ 0 .00015 °/ √hr , and if the
orientation bias is 0.06 °. The latter numbers all reflect an order of magnitude lower than
commensurate gradient error effects of 1 E/ √Hz . This report also provides detailed models for
the various error sources, as well as for the accelerations and angular rates of the aircraft and for
the gravitational signal to wavelengths as short as 1 m.
Technical Report prepared for National Imagery and Mapping Agency. Contract No. NMA202-98-1-1110. OSURF Project No. 736145.
2003-09-01T00:00:00ZJekeli, ChristopherOcean Tide Modeling in the Southern Ocean
http://hdl.handle.net/1811/78650
Ocean Tide Modeling in the Southern Ocean
Wang, Yu
Ocean tide has been observed and studied for a long time. With its role in the
complex interactions between solid earth, ocean, sea ice and the floating glacial ice
shelves, tides have been identified as one of the important causes of grounding line
migration, an essential factor to the study of ice mass balance and global sea level change.
In addition, accurate knowledge of ocean tides is needed for studies such as tidal mixing
and sea ice calving. Polar ocean tide models remain poorly understood despite of the
success of global ocean tide modeling in the deep oceans. In this thesis, a study of ocean
tide modeling in the Southern Ocean employing the empirical tide solution approach is
presented using the multiple satellite altimetry data at crossover locations.
The tidal aliasing problem in satellite altimetry is first investigated by testing the
software for two frequency searching methods using simulated and actual altimetry data
at crossover locations. Numerical experiments show that the software for the interval
method performs better than that for the global optimization frequency searching method,
by which the true original (not aliased) frequencies of the tides can be extracted from
altimetry data at crossover locations. Also, using altimetry data at crossover locations can
better reduce tidal aliasing than using along-track altimetry data.
Altimetry data at T/P and ERS-2 dual satellite crossovers for the Southern Ocean are
generated using 300 cycles of T/P data and 79 cycles of ERS-2 data. Using these data, an
empirical ocean tide solution is derived using the orthotide formulations. Different
weighting methods are tested, and the use of different weights at different locations is
adopted as our solution strategy. The empirical tide solutions have been evaluated by
comparison with several other models, including the global tide models NAO99,
TPXO.6.2 and the regional model CATS02.01. The comparison shows that our solution is
comparable with the selected models. The RSS of 8 major short-period ocean tides
between our solution using altimetry data at dual satellite crossover locations and the
selected models is 2.2 ~ 2.4 cm. And when compared with selected models in terms of
standard deviation of the sea surface height residuals, our solution shows improved
performance with a tidal power of 22 ~ 42 cm2 improvement over the selected models.
This report was prepared by Yu Wang, a graduate research associate in the Department of Civil and Environmental Engineering and Geodetic Science at the Ohio State University, under the supervision of Professor C. K. Shum. This report was supported by a grant from the National Science Foundation Office of Polar Program: OPP-0088029.; This report was also submitted to the Graduate School of the Ohio State University as a thesis in partial fulfillment of the requirements for the Master of Science degree.
2004-12-01T00:00:00ZWang, YuLocal and Regional Geoid Determination from Vector Airborne Gravimetry
http://hdl.handle.net/1811/78649
Local and Regional Geoid Determination from Vector Airborne Gravimetry
Serpas, Juan Gilberto, 1959-
The local geoid in a test area in the Canadian Rocky Mountains is computed using
airborne gravimetry data. The geoid is computed by the use of the vertical and horizontal
components (VC and HC) of the gravity disturbance vector. In addition, an attempt to
combine the three components by the use of least squares collocation is done. The
technique of using crossovers to estimate for biases and trends in the gravity signals and
the use of minimal control in the form of constraints in the crossover adjustment are
studied. Moreover, the downward continuation as well as the direct and indirect effects
due to removal and restoration of the masses are investigated. An expression for the
effect of the masses applied directly to disturbing potential is provided.
Comparison of the predicted components of the gravity disturbing vector with
control data indicates that the vertical component is better determined than the horizontal
component. The estimated accuracy for the vertical components is on the order of 4
mGal, whereas for the horizontal components it is on the order of 8 to 12 mGal.
Both geoid estimates coming from the vertical and horizontal components of the
gravity disturbance vector, computed using Hotine s and line integral, show the same
level of accuracy when compared to the Canadian geoid. Relative geoid accuracies on the
order of 3 to 7 cm for the VC geoid, and on the order of 4 to 12 cm for the HC geoid are
achieved.
The VC geoid suffers from edge effects on the results, while the HC geoid is
highly dependent on ground control. In order to alleviate the use of full ground control
for the HC geoid, the computation of the geoid at two crossing tracks is explored.
Regarding the estimation of the geoid using least squares collocation to combine
the three components of the gravity disturbance vector (3C-LSC), we observe differences
in the range of 4 to 6 cm, without including edge effects, with respect to the Canadian
geoid. Comparing the 3C-LSC results with those from the VC geoid using Hotine s
integral, the 3C-LSC are comparable and improved for some lines, in terms of standard
deviation. In general the result from the 3C-LSC are better than those from HC, by line
integral. On the other hand, the use of only the vertical component by least squares
collocation (VC-LSC) provides, in general, better results than those from 3C-LSC, and
those from the VC and HC by the use of Hotine s and line integral. We could expect
better results for the case of 3C-LSC if we are able to improve the quality of the
measured horizontal components of the gravity disturbance vector.
The application of a wave correlation filter to both HC and VC component geoid
is also explored, and promising results for the improvement of the accuracy of the
combined geoid are observed.
This report was prepared by Juan Gilberto Serpas, a graduate student, Department of
Civil and Environmental Engineering and Geodetic Science, under the supervision of
Professor Christopher Jekeli.; This research was supported under a contract with the National Imagery and Mapping
Agency (NIMA). Contract no. NMA202-98-1-1110.; This report was also submitted to the Graduate School of The Ohio State University as a
thesis in partial fulfillment of the requirements for the degree Doctor of Philosophy.
2003-10-01T00:00:00ZSerpas, Juan Gilberto, 1959-Efficient Global Gravity Determination from Satellite-to-Satellite Tracking (SST)
http://hdl.handle.net/1811/78648
Efficient Global Gravity Determination from Satellite-to-Satellite Tracking (SST)
Han, Shin-Chan
By the middle of this decade, measurements from the CHAMP (CHAllenging of Minisatellite Payload)
and GRACE (Gravity Recovery And Climate Experiment) gravity mapping satellite missions are
expected to provide a significant improvement in our knowledge of the Earth's mean gravity field and its
temporal variation. For this research, new observation equations and efficient inversion method were
developed and implemented for determination of the Earth’s global gravity field using satellite
measurements. On the basis of the energy conservation principle, in situ (on-orbit) and along track
disturbing potential and potential difference observations were computed using data from
accelerometer- and GPS receiver-equipped satellites, such as CHAMP and GRACE. The efficient
iterative inversion method provided the exact estimates as well as an approximate, but very accurate
error variance-covariance matrix of the least squares system for both satellite missions.
The global disturbing potential observable computed using 16-days of CHAMP data was used to
determine a 50´50 test gravity field solution (OSU02A) by employing a computationally efficient
inversion technique based on conjugate gradient. An evaluation of the model using independent
GPS/leveling heights and Arctic gravity data, and comparisons with existing gravity models, EGM96
and GRIM5C1, and new models, EIGEN1S and TEG4 which include CHAMP data, indicate that
OSU02A is commensurate in geoid accuracy and, like other new models, it yields some improvement
(10% better fit) in the polar region at wavelengths longer than 800 km.
The annual variation of Earth’s gravitational field was estimated from 1.5 years of CHAMP data and
compared with other solutions from satellite laser ranging (SLR) analysis. Except the second zonal and
third tesseral harmonics, others second and third degree coefficients were comparable to SLR solutions
in terms of both phase and magnitude. The annual geoid change of 1 mm would be expected mostly
due to atmosphere, continental surface water, and ocean mass redistribution. The correlation between
CHAMP and SLR solutions was 0.6~0.8 with 0.7 mm of RMS difference. Although the result should
be investigated by analyzing more data for longer time span, it indicates the significant contribution of
CHAMP SST data to the time-variable gravity study.
Considering the energy relationship between the kinetic and frictional energy of the satellite and the
gravitational potential energy, the disturbing potential difference observations can be computed from the
orbital state vector, using high-low GPS tracking data, low-low satellite-to-satellite GRACE
measurements, and data from 3-axis accelerometers. Based on the monthly GRACE simulation, the
geoid was obtained with an accuracy of a few cm and with a resolution (half wavelength) of 160 km.
However, the geoid accuracy can become worse by a factor of 6~7 because of spatial aliasing. The
approximate error covariance was found to be a very good accuracy measure of the estimated
coefficients, geoid, and gravity anomaly. The temporal gravity field, representing the monthly mean
continental water mass redistribution, was recovered in the presence of measurement noise and high
iii
frequency temporal variation. The resulting recovered temporal gravity fields have about 0.2 mm errors
in terms of geoid height with a resolution of 670 km.
It was quantified that how significant the effects due to the inherent modeling errors and temporal
aliasing caused by ocean tides, atmosphere, and ground surface water mass are on monthly mean
GRACE gravity estimates. The results are based on simulations of GRACE range-rate perturbations
due to modeling error along the orbit; and, their effects and temporal aliasing on the estimated
gravitational coefficients were analyzed by fully inverting monthly simulated GRACE data. For ocean
tides, the study based on the model difference, CSR4.0–NAO99, indicates that some residual
constituents like in S2 may cause errors 3 times larger than the measurements noise at harmonic degrees
less than 15 in the monthly mean estimates. On the other hand, residual constituents in K1, O1, and M2
are reduced by monthly averaging below the measurement noise level. For the atmosphere, the
difference in models, ECMWF–NCEP, produces errors in GRACE range-rate measurements as strong
as the measurement noise. They corrupt all recovered coefficients and introduce 30 % more error in the
global monthly geoid estimates up to maximum degree 120. However, the analysis based on daily
CDAS-1 data for continental surface water mass redistribution indicates that the daily soil moisture and
snow depth variations affect the monthly mean GRACE recovery less than the measurement noise.
This report was prepared by Shin-Chan Han, a graduate student, Department Civil and Environmental Engineering and Geodetic Science, under the supervision of Professors C. Jekeli and C.K. Shum.; This research is supported by grants from the Center for Space Research, University of Texas under a prime contract from NASA (CSR/GRACE #735367), and from NASA's Solid Earth and Natural Hazards Program (NASA/GRACE #736312).; This report was also submitted to the Graduate School of Ohio State University as a thesis in partial fulfillment of the requirements for the degree Doctor of Philosophy.
2003-09-01T00:00:00ZHan, Shin-ChanRadar Altimeter Absolute Calibration Using GPS Water Level Measurements
http://hdl.handle.net/1811/78647
Radar Altimeter Absolute Calibration Using GPS Water Level Measurements
Cheng, Kai-chien, 1969-
Recent studies of using long-term island and coastal tide gauges (over 60 years) indicate
that the global sea level rise is at a rate of 1.8 to 1.9 ± 0.1 mm/year (e.g., Douglas 1997, 1991;
Trupin and Wahr, 1990; Warrick and Oerlemans, 1990). Satellite radar altimetry has evolved
into a tool for synoptic observation of the global (±81.5° latitude) oceanic phenomena with
unprecedented accuracy (several cm in sea surface height) and with a temporal resolution of 1-2
weeks and a spatial resolution of 50 km. Its accuracy, global coverage, and temporal
resolution enable its use in studies of global sea level changes. With accurate links among
different satellite radar altimeters, a decadal (~ 15 years) altimeter sea surface height (ssh)
measurements can be obtained. However, the limitations of using altimeters to measure sea
level include inadequate knowledge of the instrumental biases and their potential drifts of each
individual radar altimeter.
The inherent requirements to enable the use of radar altimeters to measure global ssh
include knowledge of the altimeter biases to within 1-cm accuracy and their drifts to less than 1
mm per year. The mechanism is to conduct the absolute radar altimeter calibration with the
eventual goal to obtain the knowledge of its bias and drift with sufficient accuracy for sea level
studies. The goal of the radar altimeter absolute calibration is to determine the altimeter bias
and drift by comparing the altimeter-measured ssh with the accurate ground truth, often referred
to as in situ data sets. However, both systems contain different error sources and it is
necessary to formulate a closure equation and solved for the altimeter bias and drift with least
squares.
In this paper, a GPS buoy campaign in Lake Michigan was conducted by the
Laboratory for Space Geodesy and Remote Sensing Research of the Department of Civil and
Environmental Engineering and Geodetic Science, at the Ohio State University (CEEGS/OSU)
in cooperation with the National Geodetic Survey, National Oceanic and Atmospheric
Administration (NGS/NOAA) from March 20 to 24, 1999. The lake was chosen because of
the relatively calm water conditions such as waves and wind compared with oceans. 1-Hz
kinematic data obtained from a GPS buoy and the lake level record from 1993 to 1999
collected by the Holland West tide gauge were used in this study as the in situ data sets for the
absolute calibration of TOPEX/POSEIDON Side A (TSA) and Side B (TSB).
The GPS buoy and the Holland West tide gauge have not yet met the required accuracy
of absolute radar altimeter calibration. The geoid gradient is the most dominant error source
among others and it should be carefully avoided. The T/P bias and drift estimations in this
study are inaccurate because of the large geoid gradient in using GPS buoy data and the 10-cm
discrepancy in using the Holland West tide gauge data. However, it can be anticipated that the
accuracy will be improved with more data in the future and also with more data collected by
other calibration sites worldwide.
The report was prepared by Kai-chien Cheng, a graduate research associate in the Department of Civil and Environmental Engineering and Geodetic Science, at the Ohio State University, under the supervision of Professor C. K. Shum. This report was supported in part by the National Oceanographic Partnership Program Grant (Dynalysis of Princeton #865618) and National Aeronautics and Space Administration TOPEX/POSEIDON Extended Mission Grant (NAG5-6910/JPL961462), National Aeronautics and Space Administration Earth Science Information Partnership CAN Grant (CIT #12024478), National Aeronautics and Space Administration Interdisplanary Science Project (NAG5-9335), National Science Foundation Digital Government Grant (EIA-0091494, and the Ohio Sea Grant Program (R/CE-5).; This report was also submitted to the Graduate School of the Ohio State University as a thesis in partial fulfillment of the requirements for the Master of Science degree.
2004-01-01T00:00:00ZCheng, Kai-chien, 1969-INS/GPS Vector Gravimetry Along Roads in Western Montana
http://hdl.handle.net/1811/78646
INS/GPS Vector Gravimetry Along Roads in Western Montana
Jekeli, Christopher; Li, Xiaopeng
This document reports on a comprehensive look at the data collected by the GPSVan (OSU,
Center for Mapping) in western Montana in April and June of 2005. The data consist of inertial
measurement unit (IMU) data, extracted from high-accuracy inertial navigation systems, and
differential GPS data that are combined to estimate the (3-D) gravity vector along the roadways
traveled by the vehicle. The key to the evaluation of these tests and to their deemed success was
the repeated runs of the traverses, rather than the existing control data in the region. The fairly
dense network of gravity data provided only some overall corroboration of the accuracy in the
vertical components of the estimates. The deflection of the vertical (DOV) data and other
independent sources of computed DOV’s provided barely some long-wavelength confirmation of
our estimates, while the repeatability of the traverses verified fine detail in the recovered
horizontal components. However, this precision was not consistent and large errors remain in
the horizontal components. The single largest detriment to our estimates was the inaccuracy in
the kinematic GPS positioning solution. Due to road overpasses and other obstructions, the GPS
solution was often degraded significantly due to the inability to solve for the cycle ambiguity.
This had a direct and demonstrable effect on the gravity estimation. When all systems were
working at peak performance, we showed better than mgal repeatability in the down component
of the gravity disturbance and standard deviation of 2-3 mgal with respect to the interpolated
control data. No attempt was made in this first analysis to solve for biases and linear trends, nor
to take advantage of the multiple traverses to arrive at final along-track gravity disturbance
estimates. Three essential conclusions were obtained from our analysis: 1) GPS solutions must
be improved, e.g., using INS to help recover the cycle ambiguity after a GPS outage; 2) more
direct, along-track control data are necessary, particularly in the horizontal components, to obtain
a meaningful assessment of the vector gravimetry capability of the system; and 3) an operational
system would clearly benefit from redundancy in instrumentation in order to imitate and take
advantage of multiple traverses along each surveyed road.
The first chapter summarizes the instrument setup, the survey routes, the data collected, and the
control data available. The second chapter briefly reviews the techniques used to obtain the
gravity vector estimates, relying heavily on previous publications and reports. Results of
applying these techniques to the data are shown in Chapter 3; followed by the concluding chapter
with comments and analyses, and an outlook toward further data processing.
This project is supported under a contract with the National Geospatial-Intelligence Agency,
contract no. NMA- NMA401-02-1-2005; and this report serves as an Interim Technical Report
for the project.
2006-01-01T00:00:00ZJekeli, ChristopherLi, XiaopengImplementation of Parallel Least-Squares Alorithms for Gravity Field Estimation
http://hdl.handle.net/1811/78645
Implementation of Parallel Least-Squares Alorithms for Gravity Field Estimation
Xie, Jing
NASA/GFZ’s Gravity Recovery and Climate Experiment (GRACE) twin-satellite
mission, launched in 2002 for a five-year nominal mission, has provided accurate
scientific products which help scientists gain new insights on climate signals which
manifest as temporal variations of the Earth’s gravity field. This satellite mission also
presents a significant computational challenge to analyze the large amount of data
collected to solve a massive geophysical inverse problem every month. This paper
focuses on applying parallel (primarily distributed) computing techniques capable of
rigorously inverting monthly geopotential coefficients using GRACE data. The gravity
solution is based on the energy conservation approach which established a linear
relationship between the in-situ geopotential difference of two satellites and the position
and velocity vectors using the high-low (GPS to GRACE) and the low-low (GRACE
spacecrafts) satellite-to-satellite tracking data, and the accelerometer data from both
GRACE satellites. Both the direct or rigorous inversion and the iterative (conjugate
gradient) methods are studied. Our goal is to develop numerical algorithms and a portable
distributed-computing code, which is potentially “scalable” (i.e., keeping constant
efficiency with increased problem size and number of processors), capable of efficiently
solving the GRACE problem and also applicable to other generalized large geophysical
inverse problems.
Typical monthly GRACE gravity solutions require solving spherical harmonic
coefficients complete to degree 120 (14,637 parameters) and other nuisance parameters.
The accumulation of the 259,200 monthly low-low GRACE observations (with 0.1 Hz
sampling rate) to normal equations matrix needs more than 55 trillion floating point
operations (FLOPs) and ~1.7 GB central memory to store it. Its inversion adds ~1 trillion
FLOPs. To circumvent this huge computational challenge, we use a 16 nodes SGI 750
cluster system with 32 733 MHz Itanium processors to test our algorithm. We choose the
object-oriented Parallel Linear Algebra Package (PLAPACK) as the main tool and
Message Passing Interface (MPI) as the underlying communication layer to build the
parallel code. MPI parallel I/O technique is also implemented to increase the speed of
transferring data between hard drive and memory. Furthermore, we optimize both the
serial and parallel codes by carefully analyzing the cost of the numerical operations, fully
exploiting the power of the Itanium architecture and utilizing highly optimized numerical
libraries.
For direct inversion, we tested the implementations of the Normal equations
Matrix Accumulation (NMA) method, that computes the design as well as normal
iii
equations matrix locally and accumulates them to global objects afterwards, and the
Design Matrix Accumulation (DMA) approach, which forms small-size design matrices
locally first and transfers them to global scale by matrix-matrix multiplication to obtain a
global normal equations matrix. The creation of the normal equations matrix takes the
majority of the entire wall clock time. Our preliminary results indicate that the NMA
method is very fast but at present cannot be used to estimate extremely high degree and
order coefficients due to the lack of central memory. The DMA method can solve for all
geopotential coefficients complete to spherical harmonic degree 120 in roughly 30
minutes using 24 CPUs. The serial implementation of the direct inverse method takes
about 7.5 hours for the same inversion problem using the same but only one processor.
In the realization of the conjugate gradient method on the distributed platform, the
preconditioner is chosen as the block diagonal part of the normal equations matrix. The
approximate computation of the variance-covariance matrix for the solution is also
implemented. With significantly less arithmetic operations and memory usage, the
conjugate gradient method only spends approximately 8 minutes (wall clock time) to
solve for the gravity field coefficients up to degree 120 using 24 CPUs after 21 iterations,
while the serial code runs roughly 3.5 hours to achieve the same results on a single
processor.
Both the direct inversion method and the iterative method give good estimates of
the unknown geopotential coefficients. In this sense, the iteration approach is better for
the much shorter running time, but only an approximation of the estimated variancecovariance
matrix is provided. Scalability of the direct and iterative method is also
analyzed in this study. Numerical results show that the NMA method and the conjugate
gradient method achieve good scalability in our simulation. While the DMA method is
not as scalable as the other two for smaller problem sizes, its efficiency improves
gradually with the increase of problem sizes and processor numbers. The developed
codes are potentially transportable across different computer platforms and applicable to
other generalized large geophysical inverse problems.
This report was prepared by Jing Xie, a graduate research associate in the Department of Civil and Environmental Engineering and Geodetic Science at the Ohio State University, under the supervision of Professor C. K. Shum. This research was partially supported by grants from NSF Earth Sciences program: EAR-0327633, NASA Office of Earth Science program: NNG04GF01G and NNG04GN19G.; This report was also submitted to the Graduate School of the Ohio State University as a thesis in partial fulfillment of the requirements for the Master of Science degree.
2005-03-01T00:00:00ZXie, JingAnalysis of Water Level Measurements Using GPS
http://hdl.handle.net/1811/78644
Analysis of Water Level Measurements Using GPS
Cheng, Kai-chien, 1969-
Accurate knowledge about sea level and its change is essential to humanity because a
large proportion of the Earth's population lives in coastal regions. This study discusses the
existing techniques for sea level measurements, including the use of different types of gauges
(e.g., water level gauge or tide gauge, and bottom pressure gauge), as well as GPS and satellite
altimetry. The GPS water level measurements from a buoy or a vessel are presented and utilized
in this study along with other techniques to collect ellipsoidal, geocentric sea surface height
measurements for various studies that help improve our knowledge about sea level and its
change.
An operational technique of using GPS water level measurement is proposed in this
study. The limitation and an upper bound accuracy of the kinematic (epoch-by-epoch)
positioning in terms of baseline length are discussed. A set of GPS data in Lake Erie, including
buoy data as well as a local GPS network on land, are used to provide the numerical results.
Three main applications of using the GPS water level measurements are presented in this
study. They are integration of various data sources in the coastal, satellite radar calibration, and
GPS hydrology. The objective of these applications is to demonstrate the potential of the GPS
technique in collecting water level measurements. The use of GPS measurements is also
highlighted in connection with the improvement that they may bring to various techniques such
as the use of coastal water level gauge and bottom pressure gauge, and satellite altimetry.
The water level gauges are the traditional tools to collect water level data in the coastal
areas. A bottom pressure gauge, on the other hand, is deployed away from the shore that senses
pressure change in order to infer sea surface variation in terms of depth. Both types of gauges
provide only relative measurements, and the land, where they are installed, is subject to the local
vertical land movement. In order to take advantage of the large amount of gauge records, a GPS
buoy/vessel occupation can be made to link their relative measurements to the global reference
frame. This facilitates the integration use of the gauge records to the satellite measurements from
altimeters as well as from the GPS technique.
Since studies of global sea level rise using satellite altimetry examine the signal whose
magnitude is about 1–2 mm/year, the constant altimeter range bias and the drift should be
calibrated and accounted for with the calibration sites around the world. In this study, two
calibration sites―the Lake Erie Calibration Site and the South Pacific Calibration Site―were
established to support such a global effort for altimeter calibration. The Lake Erie Site uses a
coastal water level gauge off the satellite track by 20 km and still produces comparable results
ii
compared with others. The establishment of both sites will be address and the instruments for
water level measurements involved are: GPS buoys, vessels, satellite altimeters, coastal water
level gauge, and bottom pressure gauge.
The GPS water level measurements were also made to provide the river stage height in
the Branco River, a tributary of the Amazon River. The stage height along the river is surveyed
with a GPS ship. The stage gradient, which is the primary information for quantify sedimentation
of the river, is estimated from the GPS ship data. The standard deviation is better than ±0.4
cm/km, which is consistent with other studies in this area.
This study discusses three applications of using GPS water level measurements. They
have shown the capabilities of the GPS technique on buoys or vessels to interact with other
techniques for making accurate water level measurements. With the water impacts humanity,
such measurements have proven to be valuable for better understanding for the coastal
environment.
The Report describes the PhD dissertation research completed by Kai-chien Cheng in September, 2005, supervised by Professor C.K. Shum and other PhD Committee Members including Profs. Michael Bevis, Stéphane Calmant, and Burkhard Schaffrin.; This research is supported by grants from National Aeronautics and Space Administration's Ocean, Ice and Climate Program (NNG04GA53G and NNG04GN19G), Earth Science Enterprise Program (NGA5-12585) and Earth Observing System Interdisciplinary Science Program (NAG5-9335), National Science Foundation's Information and Intelligent Systems Digital Government Program (IIS-0091494), and the Ohio Sea Grant (No. R/CE-5) via funding from the National Oceanic and Atmospheric Administration.
2005-11-01T00:00:00ZCheng, Kai-chien, 1969-Determination and Characterization of 20th Century Global Sea Level Rise
http://hdl.handle.net/1811/78643
Determination and Characterization of 20th Century Global Sea Level Rise
Kuo, Chung-Yen, 1974-
Sea level rise has been widely recognized as a measurable signal and as one of the
consequences of a possible anthropogenic (human-induced) effect on global climate change. The
small rate of sea level rise, 1–2 mm/yr during the last century [Church et al., 2001, Chapter 11,
Changes in Sea Level, in the Third Assessment Report (TAR) of the Intergovernmental Panel for
Climate Change (IPCC), Working Group I, Houghton et al., 2001], could only be partially
explained by a number of competing geophysical processes, each of which is a complex process
within the Earth-atmosphere-ocean-cryosphere-hydrosphere system. In particular, the observed
20th Century sea level rise rate of 1.84±0.35 mm/yr [Douglas, 2001; Peltier, 2001] could not
explain up to one half of the predicted 20th Century global sea level rise based on the IPCC TAR
estimate of 1.1 mm/yr (0.6 mm/yr of melted water from ice sheets and glaciers, and 0.5 mm/yr
from the steric effect in the ocean) [Church et al., 2001] and remains an enigma [Munk, 2002].
The quest to resolve the controversy [Meier and Wahr, 2002] and to further understand sea level
change [Chao et al., 2002] is well underway including efforts being conducted during the current
IPCC Fourth Assessment Report (FAR), 2003–2007.
In this study, we provide a determination of the 20th Century (1900–2002) global sea
level rise, the associated error budgets, and the quantifications or characterization of various
geophysical sources of the observed sea level rise, using data and geophysical models. We
analyzed significant geographical variations of global sea level change including those caused by
the steric component (heat and salinity) in the ocean, and sea level redistribution resulting from
ice sheets and glacier melting in consequence of self gravitation, and the effects of glacial
isostatic adjustment (GIA) since the Pleistocene affecting sea level signals in the observations. In
particular, relative sea level data from up to 651 global long-term (longest record is 150 years)
tide gauges from Permanent Service for Mean Sea Level (PSMSL) and other sources, and
geocentric sea level data from multiple satellite radar altimetry (1985–2005) have been used to
determine and characterize the 20th Century global sea level rise. Altimeter and selected tide
gauge data have been used for sea level determination, accounting for relative biases between
different altimeters and offsets between the tide gauges, effects of thermosteric sea level
variations, vertical motions affecting tide gauge sea level measurements, sea level redistribution
due to ice melt resulting from self gravitation, and barotropic ocean response due to atmospheric
forcing. This study is also characterized by the role of the polar ocean in the global sea level
study and addressing the question whether there is a detectable acceleration of sea level rise
during the last decade. Vertical motions have been estimated by combining geocentric sea level
measurements from satellite altimetry (TOPEX/POSEIDON, T/P) and long-term relative (crustfixed)
sea level records from global tide gauges using the Gauss-Markov model with stochastic
constraints. The study provided a demonstration of improved vertical motion solutions in semienclosed
seas and lakes, including Fennoscandia and the Great Lakes region, showing excellent
agreement with independent GPS observed radial velocities, or with predictions from GIA
models. In general, the estimated uncertainty of the observed vertical motion is <0.5 mm/yr,
significantly better than other studies. Finally, improved algorithms to account for nonlinear
vertical motions caused by other geodynamic processes than GIA including post-seismic
deformations, have been developed and applied to tectonically active regions such as Alaska and
compared well with GPS velocities and other studies. This novel technique could potentially
provide improved vertical motion globally where long-term tide gauge records exist.
The thermosteric sea level trend of the upper (0–500 m) layers of the ocean accounts for
about 70% of the variations when deeper ocean (0–3000 m) is considered using the World Ocean
Altas 2001 (WOA01). The estimated global thermosteric sea level trend of 0.33 mm/yr (0–500 m)
and 0.43 mm/yr (0–3000 m) agrees well with other studies [Levitus et al., 2005; Ishii et al.,
2005]. A detailed analysis using in situ temperature and salinity data from the Ocean Station
Data (OSD), in the Eastern Pacific (where the in situ data are more abundant) is conducted to
assess the contributions of respective roles of thermal and salinity effects on sea level changes.
The analysis indicates that the estimated thermosteric sea level trends integrated from 0–500 m
and from 0–1000 m depths using OSD are almost identical at ~0.64 mm/yr in the Eastern Pacific
ocean. In this region, the halosteric (salinity) sea level trend is small, at the 0.04 mm/yr level for
both the 0–500 m and 0–1000 m cases. The result is consistent with Miller and Douglas [2004],
which analyzed data from selected oceanic regions including the Eastern Pacific, Eastern
Atlantic, and Caribbean. In general, observed sea level (from tide gauges or altimetry) and
thermosteric sea level are highly correlated except in regions of high mesoscale variability. A
detailed comparison of the estimated thermosteric sea level trends with the observed sea level
trend near proximities of global tide gauges (~50 years, 1950–2004, up to 597 sites) indicates
that the thermosteric sea level trends account for 36% and 68% of the observed global sea level
trend, for the 0–500 m and 0–3000 m cases, respectively, indicating the importance of deeper
ocean thermal effects. Comparison of global thermosteric sea level trend with altimetry observed
sea level trends indicates that the results are dependant on data spans used: thermosteric sea level
trend (0–500 m) accounts for over 81% of total observed sea level using T/P altimetry during
1993–2003, while accounts for only 26% when data span extends to 1985–2005 using multiple
altimetry. The primary reason is the presence of long-period oceanic variability (interannual,
decadal or longer) in trend estimates causing inconsistent conclusions. Comparison of
thermosteric sea level trend with sea level observed by selected tide gauges reaches different
conclusions depending on the choice and number of tide gauges (1955–1996) used: thermosteric
sea level trend (0–3000 m) accounts for over 88% of the observed sea level from 27 tide gauges,
while it accounts for only 38% if 515 tide gauges are used.
A study using temporal gravity field measurements observed by the Gravity Recovery
and Climate Experiment (GRACE) in terms of oceanic mass variations at long-wavelength (>800
km) and monthly sampling demonstrates its potential use, when combined with satellite altimetry,
to improve steric sea level trend estimates over the Southern Ocean, where the OSD is extremely
sparse or non-existent.
The global (covering within ±81°.5 latitude) sea level trend (corrected for IB) observed
by multiple satellite altimetry (GEOSAT, ERS-1, T/P, ERS-2, GFO, JASON and ENVISAT,
data covering 1985–2002 except for 1988–1991), and by ~530 tide gauges are 2.8±0.5 mm/yr
(corrected for GIA geoid effect) and 2.7±0.4 mm/yr, respectively, indicating excellent agreement.
The observed sea level trend using multiple satellite altimeters covering 1985–2004 is 2.9±0.5
mm/yr, although the data span is still too short to yield reliable trend estimates. The multiple
altimeter data sets allow coverage in the polar oceans, as opposed to only using data from T/P
(coverage within ±66°), and have a longer data span (~19 years versus 10 year for T/P). The data
have been calibrated against each other and with tide gauge data for robust determination of their
relative biases with respect to T/P. The 20th Century (1900–2002) global sea level trend
determined from 651 tide gauge stations is 1.6±0.4 mm/yr, while the trend estimate for the last
50 years (1948–2002) is similar, but used 620 tide gauges. The atmospheric effect via the
barotropic IB correction (globally averaged at ~0.11 mm/yr and over 50 years at global tide
gauge locations) to sea level (tide gauges or altimetry) observations is warranted as it reduces
variances of the data and causes improved agreements between altimetry and tide gauge sea level
data. The multiple altimeter data allows sea level studies in the Arctic and Sub-Arctic Oceans,
which have been less studied. The sea level trend in the Sub-Arctic Ocean, bounded by latitude
55°N–82°N and longitude 315°E–60°E, is estimated at 1.8 mm/yr using primarily ERS-1 and
ERS-2 altimetry, which have been calibrated by TOPEX/POSEIDON in the lower latitude (±66°
latitude) oceans and remove constant offsets and a latitude-dependent bias. The sea level trend
can mostly be explained by atmospheric forcing and thermosteric effects. In the Arctic Ocean,
sea level trend during 1948–2002 is estimated at 1.9–2.0 mm/yr using tide gauges after
correction of land motion using different GIA models. After applying the inverted barometric
(IB) correction, the sea level rate reduces to 1.5–1.6 mm/yr, indicating that the barotropic
response of the ocean contributes significantly more to the sea level trend in the Arctic Ocean
than for the global ocean.
In an attempt to estimate the global sea level trend and to quantify some of the known
contributions in the 20th Century using data from sparsely distributed long-term tide gauges and
~20 years of satellite altimetry, we assume that the spatial patterns of sea level trends from melt
water sources (Antarctica, Greenland and mountain glaciers), the thermosteric effects, and geoid
change and land uplift due to GIA are known, while their magnitudes are unknown. We ignored
contributions from fresh water imbalances due to continental hydrological processes [Milly et al.,
2003; Ngo-Duc et al., 2005], human-impoundment of water in reservoirs or lakes [Chao et al.,
1994; Sahagian et al., 1994] and other effects, including permafrost melt [Zhang et al., 2005].
Using two different estimators, the Weighted Least Squares (WLS) and the Elementwise-
Weighted Total Least Squares (EW-TLS), and using long-term tide gauges and satellite altimetry,
we estimate the respective contributions of each of the sources to the global sea level rise. The
EW-TLS technique is found to be more stable while the WLS technique produces solutions with
larger error of ~0.3 mm/yr in a simulated study. The EW-TLS solution yields the estimated 20th
Century (1900–2002) sea level trend contributions by melt water from the Antarctic and
Greenland ice sheets and mountain glaciers to be 0.80±0.14 mm/yr, 0.56±0.12 mm/yr and
0.31±0.08 mm/yr, respectively; the estimated contribution from the thermosteric effect is
0.06±0.04 mm/yr, and the estimated scale of the GIA (ICE-4G) model to correct tide gauge sea
level data is 1.27±0.08, indicating that the correction should be higher by 27%. The resulting
20th Century (1900–2002) globally averaged sea level trend is estimated to be 1.73±0.42 mm/yr
(95% confidence or 2s) after summing the above forcing factors. The estimate of the resulting
last 50 year (1948–2002) global sea level trend is 1.74±0.48 mm/yr (95% confidence or 2s).
Finally, we address the issue of whether the sea level trend acceleration is detectable. An
analysis indicates that the minimum data span to obtain a stable rate of sea level trend from 27
selected tide gauges [Douglas, 2001] is 20 years or more, while one should use a 30-year or
longer data span to derive a stable thermosteric sea level trend from WOA01. It is concluded
that, with 95% confidence, there is no statistically significant evidence of sea level acceleration
during 1900–2000 from tide gauge data, and during 1950-2000 from thermosteric data.
The estimated 20th Century global sea level rise is compared with the more recently
estimated total geophysical effects contributing to global sea level rise of 1.41–1.53 mm/yr,
which include the sum of the steric effect (~0.4 mm/yr) [e.g., Levitus et al., 2005; Antonov et al.,
2005], mountain glacier melting (~0.51 mm/yr) [e.g., Arendt et al., 2002; Dyurgerov and Meier,
2005; Raper and Braithwaite, 2005; and others], ice sheet mass imbalance (~0.45 mm/yr) [e.g.,
Krabill et al., 2004, Rignot et al., 2005, Thomas et al., 2005], hydrological imbalance (~0.0–0.12
mm/yr) [Milly et al., 2003, Ngo-Duc et al., 2005], and anthropogenic effect (~0.05 mm/yr) [Dork
Sahagian, pers. comm.]. This study (1.74±0.48 mm/yr) reduces the existing discrepancy [Chruch
et al., 2001] to 0.21–0.33 mm/yr between the total observed and predicted contributions to 20th
Century global sea level rise. However, the estimated individual contributions of the ice sheet
imbalance and the oceanic steric effect remain significantly different from the current results
from observations, and not much progress has been made since the last IPCC study on the effect
of hydrologic imbalance and anthropogenic causes. Much future work remains to improve our
understanding of the complex processes governing global sea level changes.
The Report describes the PhD dissertation research completed by Chung-Yen Kuo on November, 2005, supervised by C.K. Shum and other PhD Committee Members including Douglas Alsdorf, Michael Bevis, Laury Miller, and Yuchan Yi.; This research is supported by grants from NOAA under NA16RG2252, NA86RG0053 (R/CE-5) and NA030AR4170060 (R/CE-8), and NASA's Ocean and Ice, Cryosphere, Physical Oceanography and Interdisciplinary Science Programs (NNG04GA53G, NAG5-9335, NAG5-12585, JPL- 1265252).
2006-04-01T00:00:00ZKuo, Chung-Yen, 1974-GPS Radio Occultation and the Role of Atmospheric Pressure on Spaceborne Gravity Estimation Over Antarctica
http://hdl.handle.net/1811/78642
GPS Radio Occultation and the Role of Atmospheric Pressure on Spaceborne Gravity Estimation Over Antarctica
Ge, Shengjie, 1973-
Dedicated satellite gravity missions are anticipated to significantly improve the
current knowledge of the Earth’s mean gravity field and its time variable part–climate
sensitive gravity signals. They could be measured by the Gravity Recovery and
Climate Experiment (GRACE) twin-satellite with sub-centimeter accuracy in terms
of column of water movement near the Earth’s surface with a spatial resolution of
several hundred kilometers or larger, and a temporal resolution of one month or weeks.
To properly recover the time variable gravity signals from space, the gravity measurements
require the atmospheric pressure contribution to be accurately modeled
and removed. The sparse coverage of measurements makes the weather products
less accurate in the southern hemisphere, especially over the Southern Ocean and
Antarctica. The asynoptic observation from GPS radio occultation could achieve
dense spatial coverage even in remote regions. In this research, we investigate the
potential use of GPS occultation to improve the pressure modeling over Antarctica.
Atmospheric pressure profiles are retrieved and validated against ECMWF, NCEP
and radiosonde observations. Our results show that occultation can provide compatible
observations especially in the upper atmosphere. Large standard deviations and
biases are found near the ground and in the Antarctic region. GPS occultation in the
polar regions is less affected by multipath problem and can penetrate down near the
surface. Through an experiment using a 1-D variational (1DVar) approach, we show
that the high vertical accuracy of GPS occultation can be propagated down to reduce
the uncertainty of surface pressure, indicating that GPS occultation can be expected
to have positive impact on the pressure modeling over data-sparse areas after obtaining
adequate number of observations (e.g., from Constellation Observing System
for Meteorology, Ionosphere & Climate (COSMIC)). We also find that the retrieved
profiles could be different due to various assumptions and retrieval algorithms.
Pressure uncertainty degrades the GRACE recovered gravity change. We study
the uncertainty of pressure modeling on various temporal scales. Global analysis
models show large differences in the Antarctic region. The surface topography
may introduce additional biases if it is not well treated. The atmospheric tides are
non-negligible and need to be properly considered. The real magnitude of the mismodeled
and un-modeled errors in the analysis is hard to evaluate, especially in
Antarctica. We simulate the errors sensitive to GRACE using the differences between
two global analysis models. Most of the very long wavelength errors are well
ii
captured by GRACE. Their changes in the form of short-period variation increase the
errors of the middle to high degree spherical harmonic coefficients. After de-aliasing,
middle to high degree coefficients are noticeably improved. The Inverted Barometer
(IB) assumption decreases the amplitude of the aliasing error, and the pattern of
the RMS difference is slightly changed over land by neglecting the large variations in
the Southern Ocean. Our result using more recent ECMWF and NCEP operational
analyses shows reduced aliasing effects, which indicates that two models are becoming
increasingly close to each other. The model correlation and IB assumption may
underestimate the true aliasing error.
The analysis models are validated against the unevenly distributed Automatic
Weather Station (AWS) surface pressure observations on the Antarctic continent.
Spectral analysis shows that 6-hour analyzed model data can capture most of the
power in pressure variations. ECMWF exhibits a much better agreement with AWS
than NCEP reanalysis does. Large biases still exist due to the uncertainties of the
station elevations. The comparison statistics show strong correlations with the topography
with lower standard deviation values in the interior and higher standard deviation
values around the coastal area. This result contradicts the distribution derived
from the difference between two analysis models, which exhibits large difference in
the interior of Antarctica.
We also investigate the influences of different algorithms and assumptions of 2-
D or 3-D atmospheric structures on the GRACE atmospheric de-aliasing product.
Air density derived from the hydrostatic equation and the equation of state gives
slightly different results, and the difference is above the expected GRACE sensitivity.
We compare our results with the GRACE atmospheric de-aliasing product and
find that the difference is almost below the GRACE sensitivity, although there are
differences in the algorithms and we use a relatively low resolution model. We also
find that the difference between 3-D hydrostatic formulation and 2-D algorithm is
below the GRACE sensitivity. We discover that the atmospheric structure and latitudinal
variations of gravity are largely compensated by removing their respective
long-term means. Consequently, the 2-D method can greatly reduce the requirements
for computational load and data storage. Removing the mean field does not help to
reduce the discrepancies between ECMWF and NCEP. If the computational burden
is not a concern, using our improved 3-D algorithm can bring a better result. After
the full operation of the COSMIC satellites, some major improvement of the pressure
modeling over Antarctica is anticipated. A reprocessing of the GRACE data using
an improved pressure model could bring us better gravity solutions.
This report was prepared by Shengjie Ge, a graduate research associate in the Geodetic Science and surveying program of the Department of Geological Science at the Ohio State University, under the supervision of Professor C. K. Shum.; This study was partially supported by grants from NASA Interdisciplinary Science Program NAG5-9518, and National Science National Space Weather Program ATM- 0418844.; This report was also submitted to the Graduate School of the Ohio State University as a dissertation in partial fulfillment of the requirements for the Ph.D. degree.
2006-07-01T00:00:00ZGe, Shengjie, 1973-Investigations Into Green’s Function as Inversion-Free Solution of the Kriging Equation, With Geodetic Applications
http://hdl.handle.net/1811/78641
Investigations Into Green’s Function as Inversion-Free Solution of the Kriging Equation, With Geodetic Applications
Cheng, Ching-Chung, 1968-
Statistical interpolation has been proven to be a legitimate and efficient approach
for data processing in the field of geodetic and geophysical sciences. Pursuing the
minimization of the mean squared prediction error, the technique, known as Kriging or
least-squares collocation, is able to densify, respectively filter a spatially and/or
temporally referenced dataset, provided that its associated covariance model is given or
estimated in advance. The involvement of the covariance matrix which to some extent
reflects the physical behavior of the underlying process may, however, potentially lead to
an ill-conditioned situation when the data are observed at a relatively high sampling rate.
A new perspective, interpreting the Kriging equation in the continuous sense, is
therefore proposed in this research so that, instead of matrix terms, a convolution
equation is set up for the Green’s function where the covariance function is preserved in
its analytic form. Two methods to approximate the solution of such a convolution
equation are employed: One transforms the unknown Green’s function into a series
consisting of a linear combination of (partial) derivatives of the covariance function so
that the approximation of the Green’s function can be determined through a term-by-term
approach; the other one manipulates the convolution equation in the spectral domain
where the inversion can be treated within the space of real number.
The proposed approach has been applied to various covariance models, especially
several more recently established spatial-temporal models which have attracted
increasing interests for geophysical applications. Examples from geodetic science include
the cases of data fusion and terrain profile monitoring; although based on simulated data,
the demonstration of this innovative approach shows great potential.
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University.
2004-12-01T00:00:00ZCheng, Ching-Chung, 1968-On Improving the Accuracy and Reliability of GPS/INS-Based Direct Sensor Georeferencing
http://hdl.handle.net/1811/78640
On Improving the Accuracy and Reliability of GPS/INS-Based Direct Sensor Georeferencing
Yi, Yudan, 1974-
Due to the complementary error characteristics of the Global Positioning System
(GPS) and Inertial Navigation System (INS), their integration has become a core
positioning component, providing high-accuracy direct sensor georeferencing for
multi-sensor mobile mapping systems. Despite significant progress over the last decade,
there is still a room for improvements of the georeferencing performance using
specialized algorithmic approaches. The techniques considered in this dissertation include:
(1) improved single-epoch GPS positioning method supporting network mode, as
compared to the traditional real-time kinematic techniques using on-the-fly ambiguity
resolution in a single-baseline mode; (2) customized random error modeling of inertial
sensors; (3) wavelet-based signal denoising, specially for low-accuracy high-noise
Micro-Electro-Mechanical Systems (MEMS) inertial sensors; (4) nonlinear filters,
namely the Unscented Kalman Filter (UKF) and the Particle Filter (PF), proposed as
alternatives to the commonly used traditional Extended Kalman Filter (EKF).
The network-based single-epoch positioning technique offers a better way to
calibrate the inertial sensor, and then to achieve a fast, reliable and accurate navigation
solution. Such an implementation provides a centimeter-level positioning accuracy
independently on the baseline length. The advanced sensor error identification using the
Allan Variance and Power Spectral Density (PSD) methods, combined with a
wavelet-based signal de-noising technique, assures reliable and better description of the
error characteristics, customized for each inertial sensor. These, in turn, lead to a more
reliable and consistent position and orientation accuracy, even for the low-cost inertial
sensors. With the aid of the wavelet de-noising technique and the customized error model,
around 30 percent positioning accuracy improvement can be found, as compared to the
solution using raw inertial measurements with the default manufacturer’s error models.
The alternative filters, UKF and PF, provide more advanced data fusion techniques and
allow the tolerance of larger initial alignment errors. They handle the unknown nonlinear
dynamics better, in comparison to EKF, resulting in a more reliable and accurate
integrated system. For the high-end inertial sensors, they provide only a slightly better
performance in terms of the tolerance to the losses of GPS lock and orientation
convergence speed, whereas the performance improvements are more pronounced for the
low-cost inertial sensors.
2007-12-01T00:00:00ZYi, Yudan, 1974-Near Real-Time Precise Orbit Determiation of Low Earth Orbit Satellites Using an Optimal GPS Triple-Differencing Technique
http://hdl.handle.net/1811/78639
Near Real-Time Precise Orbit Determiation of Low Earth Orbit Satellites Using an Optimal GPS Triple-Differencing Technique
Bae, Tae-Suk, 1972-
During the last decade, numerous Low Earth Orbit (LEO) satellites, including
TOPEX/POSEIDON, CHAMP and GRACE, have been launched for scientific purposes
at altitudes ranging from 400 km to 1300 km. Because of highly complex dynamics of
their orbits, coming from the Earth gravity field and the atmospheric drag, accurate and
fast LEO orbit determination has been a great research challenge, especially for the
lowest altitudes. To support GPS meteorology that requires an accurate orbit in near realtime,
efficient LEO orbit determination methods were developed using the tripledifferenced
GPS phase observations, as presented in this dissertation. These methods
include the kinematic, dynamic, and reduced-dynamic approach based on the wave
algorithm.
To test the developed algorithms, 24 hours of CHAMP data on February 15, 2003,
which amounts to 15 revolutions, were used for each method. The EIGEN2 geopotential
model was used with degree and order up to 120. Precise IGS orbits are used for the GPS
satellites, and 43 IGS ground tracking stations were chosen using the algorithm
developed in this study, based on the network optimization theory. The estimated orbit
solutions were compared with the published Rapid Science Orbit (RSO) and the
consistency testing was performed for the dynamic solution. In addition to the
comparison with other orbit solutions, the SLR residuals were also computed as an
independent validation of the methods presented here.
The kinematic orbit solution depends on the satellite geometry and data quality. The
absolute kinematic positioning solution, with an RMS error of ±26 meters in 3D, was
used as an initial approximation for the kinematic orbit determination. Because of the
inaccuracy of the initial approximated orbit, there is a bias up to a few hundred epochs in
the kinematic solution. This bias is effectively removed with the backward filter by fixing
the last epoch from the forward filter solution. After the forward and backward filtering,
the kinematic approach shows accuracy better than ±20 cm in 3D RMS for a half day arc
compared to the reference RSO.
The dynamic approach requires careful modeling of the atmospheric drag force
which is the most dominant nonconservative force at LEO’s altitude. In addition, the
empirical force modeling, which is similar to the stochastic process noise in the reduceddynamic
approach, absorbs most of the remaining unmodeled forces. The two frequencies
of the empirical forces, that is, once- and twice-per-revolution, are modeled in this study.
Also after a thorough testing of the most suitable size of the arc length for the
atmospheric drag parameters, the scaling factors for the drag force are estimated every
iii
hour. With this careful modeling, the dynamic solution shows an agreement within ±8 cm
in position and ±0.12 mm/s in velocity of RSO. The computation time of the dynamic
solution for the 24-hour arc is 2.5 hours on a 3 GHz PC platform.
The wave algorithm, as implemented in this study, for the LEO precise orbit
determination (POD) represents a new approach to the reduced-dynamic technique. This
approach shows a better fit to RSO for each tested segment. However, there is slightly
larger bias in its solution, thus, the overall RMS of fit is comparable to the dynamic orbit
solution. This follows from the fact that the concept of the reduced-dynamic approach is
already incorporated in the dynamic orbit determination in the form of the empirical force
modeling. Therefore, there is no room for further improvement by the process noise
modeling to take care of the unmodeled forces. A simplified force model is considered for
future study in conjunction with the wave filter approach.
The CHAMP orbit is successfully estimated in this study to support, for example,
the GPS meteorology, using a new method that is accurate as well as fast and efficient.
The applied wave algorithm shows the possibility of further improvement in the RMS of
fit as long as the bias is modeled appropriately. The hypothesis testing indicates that the
estimated dynamic solution of this study is consistent with the published RSO, thus,
further accuracy improvement cannot be expected without other types of measurements,
while its easy and time-effective implementation represents the major improvement, as
compared to the existing solutions. Also, the SLR residual test shows that the CHAMP
orbit solution estimated in this study is comparable to solutions determined by other
analysis centers, such as JPL and GFZ.
This report was prepared by Tae-Suk Bae under the supervision of Professor Dorota A. Grejner-Brzezinska, Department of Civil and Environmental Engineering and Geodetic Science, The Ohio State University. This report was originally submitted to the Graduate School of The Ohio State University in partial fulfillment of the requirements for the Ph.D. degree.; This research was supported by the grants from NASA (NASA NIP Project, OSURF #740809).
2006-11-01T00:00:00ZBae, Tae-Suk, 1972-Recovery of Terrestrial Water Storage Change from Low-Low Satellite-to-Satellite Tracking
http://hdl.handle.net/1811/78638
Recovery of Terrestrial Water Storage Change from Low-Low Satellite-to-Satellite Tracking
Chen, Yiqun, 1975-
Gravity Recovery and Climate Experiment (GRACE) spaceborne gravimetry provides a
unique opportunity for quantifying geophysical signals including terrestrial water storage
change for a wide variety of climate change and geophysical studies. The contemporary
methodology to process GRACE data for temporal gravity field solutions is based on
monthly estimates of the mean geopotential field with a spatial resolution longer than 600
km (the Level-2 or L2 data products), after appropriate Gaussian smoothing to remove
high-frequency and geographically-correlated errors. Alternative methods include the
direct processing of the GRACE low-low satellite-to-satellite tracking data over a region
of interest, leading to improved or finer spatial and temporal resolutions of the resulting
local gravity signals. The GRACE Level 1B data have been analyzed and processed to
recover continental water storage in a regional solution, by first estimating in situ Line-
Of-Sight (LOS) gravity differences simultaneously with the relative position and velocity
vectors of the twin GRACE satellites. This new approach has been validated using a
simulation study over the Amazon basin (with three different regularization methods to
stabilize the downward continuation solutions), and it is demonstrated that the method
achieves an improved spatial resolution as compared to some of the other GRACE
processing techniques, including global spherical harmonic solutions, and regional
solutions using in situ geopotential differences.
This report was prepared for and submitted to the Graduate School of the Ohio State University as a dissertation in partial fulfillment of the requirements of the Ph.D. degree.; The research is supported by grants from NSF's Collaboration in Mathematical Geociences Program (EAR0327633), NASA Earth Science programs (NNG04GN19G, NNG05GL26G, JPL 1265252), and a Shell Fellowship (July-Sept., 2007), School of Earth Sciences, The Ohio State University. Additional computing resources are provided by the Ohio Supercomputer Center.
2007-12-01T00:00:00ZChen, Yiqun, 1975-Small Anomalous Mass Detection from Airborne Gradiometry
http://hdl.handle.net/1811/78637
Small Anomalous Mass Detection from Airborne Gradiometry
Dumrongchai, Puttipol, 1970-
A new generation of gradiometer technology is currently under development
based on atom interferometry and applicable to ground and airborne mapping of
geologic or anthropogenic features with signal strength as low as a few Eötvös,
entirely embedded in noise and geological background. With high sensitivities of
future airborne gradiometers, it may be possible to detect such anomalous sources
with careful data processing. Both the detection and the estimation of parameters of
the feature can be solved as an inverse problem in potential theory. However, one
can also use methods developed in communications theory, provided one has some a
priori, possible uncertain knowledge of the feature in question. We constructed a
matched filter as well as a sophisticated estimation technique to detect and
characterize particular small mass anomalies within general geologic background
noise using individual gradient and six gradient combination measurements at low
aircraft/helicopter altitudes of ranges of 10-30m above terrain clearance. Since both
detection and estimation portions requires the inversion of large sizes of covariance
matrices, we applied an orthogonal transformation to the matrices, which become
diagonal and can then be easily inverted. In addition, the performance of the
detection and estimation procedures is quantified by standard test statistics. With
these tests, probabilities of false alarm and detection may be assigned to the detection
results. We present numerical results in different noise circumstances, for instance, a
simulation of airborne gradiometry over moderate terrain with the inclusion of
1E/ √Hz instrumental white noise. The proposed approaches are explored and
evaluated for their effectiveness in association with location, orientation, size, and
depth of a mass anomaly, and in the use of power spectral density (psd) models
versus empirical psd’s obtained from the noise backgrounds. The numerical results
show that a small anomaly, e.g., 2m x 2m x 10m, is detectable at shallow depths by
an appropriate matched filter using, not only the empirical psd’s and the gradient
component Γ33, but also the psd models and the six-gradient combination. However,
the analysis shows that a strong noise level, low spatial resolution, and unknown
depth limit the anomaly detectability. The parameter estimation performed through
an iterative least-squares process was shown to be successful in estimating locations,
orientations, and depth of the anomaly. Hypothesis testing by means of the F-test was
used to quantify the performance of the estimation process.
This report was prepared by Puttipol Dumrongchai, a graduate student, Division of Geodesy and Geospatial Science, School of Earth Sciences at the Ohio State University under the supervision of Prof. Christopher Jekeli.; This report was also submitted to the Graduate School of the Ohio State University as a dissertation in partial fulfillment of the requirements of the Ph.D. degree.
2007-03-01T00:00:00ZDumrongchai, Puttipol, 1970-DEM Generation and Ocean Tide Modeling Over Sulzberger Ice Shelf, West Antarctica, Using Synthetic Aperture Radar Interferometry
http://hdl.handle.net/1811/78636
DEM Generation and Ocean Tide Modeling Over Sulzberger Ice Shelf, West Antarctica, Using Synthetic Aperture Radar Interferometry
Baek, Sang-Ho, 1970-
The use of Synthetic Aperture Radar Interferometry (InSAR) is an effective tool
for studying the ice mass balance of polar regions and its contribution to global sea level
change. An accurate, high-resolution digital elevation model (DEM) referenced within a
well-defined terrestrial reference frame (TRF) is an inherent requirement to facilitate the
use of InSAR to conduct these studies in remote polar regions where ground control
points (GCPs) are unavailable. In this study, a digital elevation model by the Sulzberger
Bay, West Antarctica is determined by using twelve European Remote Sensing (ERS) -1
and ERS-2 tandem satellite mission synthetic aperture radar scenes and nineteen Ice,
Cloud, and land Elevation Satellite (ICESat) laser altimetry profiles. Differential
interferograms from the ERS-1/ERS-2 tandem mission SAR scenes acquired in the
austral fall of 1996 are used together with four selected ICESat laser altimetry profiles in
the austral fall of 2004 which provides GCPs, resulting in an improved geocentric 60-m
resolution DEM over the grounded ice region. The InSAR DEM is then extended to
include two ice tongues using ICESat profiles via Kriging. Fourteen additional ICESat
profiles acquired in 2003-2004 are used to assess the accuracy of the DEM. After
accounting for radar penetration depth and predicted surface changes, including effects
due to ice mass balance, solid Earth tides, and glacial isostatic adjustment, in part to
account for the eight-year data acquisition discrepancy, the resulting difference between
the DEM and ICESat profiles is -0.55 ± 5.46 m. After removing the discrepancy between
the DEM and ICESat profiles for a final combined DEM using a bicubic spline, the
overall difference is 0.05 ± 1.35 m indicating excellent consistency.
Accurate knowledge of the Antarctic ice sheet mass balance plays an important
role on the global sea level change. Ocean tides (barotropic and baroclinic) and tidal
currents cause basal melting and migration of grounding lines, which are all critical to the
accurate determination of ice sheet or ice stream mass balance. Ocean tides in the
Antarctic Ocean, especially underneath ice shelves or sea ice, are poorly known primarily
due to lack of observations with adequate resolution and knowledge of the bathymetry
and ice shelf bottom roughness. InSAR has been used to measure the ice sheet mass
iii
balance, ice topography, ice stream velocity, and the location of the grounding lines. To
properly use InSAR measurements for ice mass balance and because of their high spatial
resolution (tens of meters), knowledge of ocean tides underneath the ice shelves needs to
be accurately known and with commensurate resolution. Here two-pass differential
InSAR (DInSAR) technique is applied for tidal signal modeling underneath the
Sulzberger ice shelf, West Antarctica. The fine resolution (60-m) Digital Elevation Model
(DEM) over grounded ice and ice shelf, obtained by combining ERS-1/2 tandem InSAR
and ICESat laser altimetry, has been used to correct the topography phase from
interferograms, resulting in a more accurate time series of vertical deformation
measurements. In this study, it is demonstrated for the first time, that observable tidal
constituents can be estimated underneath an ice shelf using an InSAR time series. In
particular, it is shown that the time series of observed tidal differences from InSAR
agrees well with a number of global/regional ocean tide models such as NAO.99b,
TPXO.6.2, GOT00.2, CATS02.01, and FES2004, with the regional model, CATS02.01,
having the best agreement. The technique developed here can be applied to other ice shelf
regions where tide modeling is poor in accuracy and resolution.
This research is supported by a grant from the National Science Foundation’s
Office of Polar Program (OPP-0088029).
2006-08-01T00:00:00ZBaek, Sang-Ho, 1970-Gradient Modeling with Gravity and DEM
http://hdl.handle.net/1811/78635
Gradient Modeling with Gravity and DEM
Zhu, Lizhi
This study deals with the methods of forward gravity gradient modeling based on
gravity data and densely sampled digital elevation data and possibly other data, such as
crust density data. In this study, we develop an improved modeling of the gravity gradient
tensors and study the comprehensive process to determine gravity gradients and their
errors from real data and various models (Stokes’ integral, radial-basis spline and LSC).
Usually, the gravity gradients are modeled using digital elevation model data under
simple density assumptions. Finite element method, FFT and polyhedral methods are
analyzed in the determination of DEM-derived gravitational gradients. Here, we develop
a method to model gradients from a combination of gravity anomaly and DEM data.
Through a solution on the boundary value problem of the potential field, the gravity
anomaly data are combined consistently with the forward model of DEM to yield nine
components of the gravity gradient tensor. As a result, forward gravity gradients can be
synthesized using both geodetic and geophysical data. We use two different methods to
process gravity data. One is the regular griding method using kriging and least squares
collocation, and the other one is based on fitting splines or wavelet functions. For DEM
data, we use finite elements, polyhedra and wavelets or splines to compute the gradients.
The second Helmert condensation principle and the remove-restore technique are used to
connect DEM and gravity data in the determination of gravity gradients.
Modeling of the gradients thus, particularly at some altitude above ground, from
surface gravity anomalies is based on numerical implementations of solutions to
boundary-value problems in potential theory, such as Stokes’ integral, least-squares
collocation, and some Fourier transform methods, or even with radial-basis splines.
Modeling of this type would offer a complementary if not alternative type of support in
the validation of airborne gradiometry systems. We compare these various modeling
techniques using FTG (full tensor gradient) data by Bell Geospace and modeled gradients,
thus demonstrating techniques and principles, as well limitations and advantages in each.
The Stokes’ integral and the least-squares collocation methods are more accurate (about 3
E at altitude of 1200 m) than radial-basis splines in the determination of gravity gradient
using synthetic data. Furthermore, the comparison between the modeled data and real
data verifies that the high resolution (higher than 1 arcmin) gravity data is necessary to
validate the gradiometry survey data.
Ground and airborne gradiometer systems can be validated by analyzing the spectral
properties of modeled gradients. Also, such modeling allows the development of survey
parameters for such instrumentation and can lead to refined high frequency power
spectral density models in various applications by applying the appropriate filter.
This report was prepared for and submitted to the Graduate School of the Ohio State University as a dissertation in partial fulfillment of the requirements of the Ph.D. degree.; The work was supported with sponsorship from the National Geospatial-Intelligence Agency under contract no. NMA401-02-1-2005.
2007-06-01T00:00:00ZZhu, LizhiSatellite Radar Altimetry for Inland Hydrologic Studies
http://hdl.handle.net/1811/78634
Satellite Radar Altimetry for Inland Hydrologic Studies
Zhang, Manman
Satellite radar altimetry, which is originally designed to measure global ocean surface
height, has been applied to inland surface water hydrologic studies. We have developed a
water-detection algorithm based on statistical analysis of decadal TOPEX/POSEIDON height
measurement time series, used the backscatter coefficient to classify the inland surface
properties, and the 10-Hz (corresponding to an along track spatial resolution of 700m) radar
waveform-retracked TOPEX data, to be able to observe small (<300Km2) inland bodies of
water for hydrologic studies. We applied the algorithm to the selected study regions in
Manitoba and northwestern (SW) Ontario, Canada, Amazon River Basin, and southwestern
Taiwan. Finally we studied the application of TOPEX altimetry to the 1997 Red River flood
monitoring. For the study regions in western Manitoba, the correlation coefficient between
stage and TOPEX altimetry data in the large Lakes reaches 0.98 using the 10-Hz retracked
data, thus verifying the validity and accuracy of the satellite measurement. The importance of
the waveform retracking for the inland water applications is validated by the improvement of
the correlation coefficients from 0.34 to 0.87 before and after retracking. We detected the
bodies of water, which are otherwise missed by using the original 1-Hz data from the
Geophysical Data Records, and illustrated that a higher spatial resolution could be achieved
using the individual 10-Hz retracked data. In the Amazon River Basin, the capability of the
water-detection algorithm is compared with the use of a high water level mask generated by
SAR and other data with a spatial resolution of 100m. It is shown that the algorithm could
detect the bodies of water, which are missed by the mask primarily because that the
frequency of water fluctuation is more than twice a year at some locations. The bodies of
water detected only by the algorithm are confirmed using the detailed local hydrological
maps in 3 tested regions. The retrieved water height over the small (<300Km2) body of water
was compared with the nearby stage measurement and showed good seasonal agreement. In
the southwest Taiwan, the monthly variation of 10-Hz AGC from 1992 to 2002 were
examined, it is found that the high AGC values could be used to indicate inundated area. We
detected the annual and semi-annual variations from the 10-Hz AGC and 10-Hz retracked
water height time series, which are attributable to two rainy seasons per year in the study
region. For the study of the 1997 Red River flood, we compared the geographic distribution
of 0 σ0 before, during and after the 1997 flood and found the high 0 σ0 values (>35dB)
indicate the inundated regions. In addition, the comparison of the geographically distributed
0 σ0 during Winter, Spring, Summer and Autumn of 1997 showed that the low 0 σ values
(<10dB) indicate snow coverage. The retrieved water height measurements in the flooded
regions are compared with the nearby USGS stage measurements and showed good
agreements. The comparison of 10-Hz individual retracked measurements with the 1-Hz nonretracked
height measurements confirmed the importance of the retracked data (with higher
spatial variations) in the flood monitoring. Using 0 σ0 and the retrieved water height
measurements, we detected the 1997 flooded regions include the Red River Basin of the
North in North Dakota and in western Minnesota, the upper Mississippi River Basin in
Minnesota, the Missouri River Basin in southern North Dakota and in South Dakota. The
observed flood extents from TOPEX agree well with and complement the USGS stage gauge
records.
This research is conducted under the supervision of Dr. C.K. Shum, Professor of
Geodetic Science, School of Earth Sciences, The Ohio State University. The research
results documented in this report resulted in a PhD Dissertation. NASA and CNES
provided the TOPEX/POSEIDON (Geophysical Data Record and Sensor Data Record,
GDR and SDR) data products; LEGOS, USDA/NASA/GSFC provided high-level radar
altimetry data products; ANA Brazil, and Environment Canada provided the stage gauge
data used for this research. This research is supported by grants from NSF’s Hydrology
Program (EAR-0440007) and NGA’s NURI Program (HM1582-07-1-2024), and the
study was conducted with the objective to contribute to the Climate, Water, and Carbon
Program at The Ohio State University.
2009-03-01T00:00:00ZZhang, ManmanA Comparative Overview of Geophysical Methods
http://hdl.handle.net/1811/78633
A Comparative Overview of Geophysical Methods
Erkan, Kamil
The shallow subsurface structure of the Earth is important to understand for many economic and safety reasons. The
problem is usually difficult due to complexity of the earth’s subsurface processes especially near the surface. A
number of geophysical methods are used for this purpose using different physical characteristics of the Earth
materials. A particular geophysical method illuminates part of the problem, but a reliable solution can only be found
by combining results of different methods. In order to synthesize information from different geophysical methods, it
is important to understand their similarities and differences. The aim of this study is to correlate the basic principles
of geophysical methods side-by-side starting from fundamental equations. This study reveals that many analogies
exist among these methods both in their mathematical formulation, and sometimes, in ways they are used in the
geophysical applications.
This report was prepared with support from the Air Force Research Laboratory, under contract FA8718-07-C-0021.
2008-09-01T00:00:00ZErkan, KamilMoving Base INS/GPS Vector Gravimetry on a Land Vehicle
http://hdl.handle.net/1811/78632
Moving Base INS/GPS Vector Gravimetry on a Land Vehicle
Li, Xiaopeng, 1975-
The Inertial Navigation System and Global Positioning System (INS/GPS) system
has been extensively studied over several decades, mostly for the purpose of
navigation and kinematic position. Because the INS system is a ected by gravitation,
the integration de nitely needs gravity data in order to yield accurate results.
It is natural to reverse the problem and attempt a measurement of the gravity vector.
The gravimetric system based on INS/GPS shows good performances in the airborne
scenarios. Moving the system into a ground vehicle will help to improve the resolution
of the gravity estimates, considering its lower speed and altitude. However, the
system will face much more complicated dynamics and harsh observation conditions.
In this study, a two-stage extended Kalman lter based on processing noise adaptation
is used to x the position gaps and provide prior information of the Inertial
Measurement Units (IMU) errors. The kinematic acceleration is computed by both
the position method and the phase method. All these procedures improve the steadiness
and precision of the system. The advanced wavelet de-noising technique is employed
to further isolate the gravity disturbance from the observation errors in the
residuals of the novel Kalman lter, previously developed at the Ohio State University
(OSU). The nal precision of the gravity disturbance estimates is further improved
by correlatively ltering the repeated estimates in the frequency domain.
An intensive survey campaign was carried out to test the validities of these techniques.
Based on data analysis, the results show signi cant consistency (as good as
0.6mGal, STD) in the vertical component on the repeated traverses, and comparison
to control data indicates an accuracy of 2-3mGal (STD). However, it is also
determined that the control data, being interpolated from a database, have an accuracy
of approximately 3mGal (STD). Resolution of the estimated gravity disturbance
is about 2km, based on 180-s data smoothing and a vehicle speed averaging about
80km/hr. Large scale errors exist in the horizontal gravity estimates. Removing
these on the basis of extensive de
ection of the vertical control yields repeatability in
the horizontal components in the range of 2-15mGal (STD) and agreement with the
control at the level of 5-9mGal (STD).
ii
This report was prepared for and submitted to the Graduate School of the Ohio State University as a dissertation in partial fulfillment of the requirements for the PhD degree.; The research was supported by a grant from the National Geospatia l Intelligence Agency (NMA202-98-1-1110) under the NGA University Research Initiative program. Additional assistance came from the Center for Mapping at the Ohio State University in the form of GPSVan survey and data analysis support.
2007-12-01T00:00:00ZLi, Xiaopeng, 1975-Geodesy in Antarctica: A Pilot Study Based on the TAMDEF GPS Network, Victoria Land, Antarctica
http://hdl.handle.net/1811/78631
Geodesy in Antarctica: A Pilot Study Based on the TAMDEF GPS Network, Victoria Land, Antarctica
Vázquez Becerra, Guadalupe Esteban, 1973-
The objective of the research presented in this report is a combination of practical
and theoretical problems to investigate unique aspects of GPS (Global Positioning
System) geodesy in Antarctica. This is derived from a complete analysis of a GPS
network called TAMDEF (Trans Antarctic Mountains Deformation), located in Victoria
Land, Antarctica. In order to permit access to the International Terrestrial Reference
Frame (ITRF), the McMurdo (MCM4) IGS (The International GNSS Service for
Geodynamics, formerly the International GPS Service) site was adopted as part of the
TAMDEF network. The following scientific achievements obtained from the cited
analysis will be discussed as follows:
(1) The GPS data processing for the TAMDEF network relied on the PAGES
(Program for Adjustment of GPS Ephemerides) software that uses the double-differenced
iono-free linear combination, which helps removing partial of bias (mm level) in the final
positioning. (2) To validate the use of different antenna types in TAMDEF, an antenna
testing experiment was conducted using the National Geodetic Survey (NGS) antenna
calibration data, appropriate for each antenna type. Sub-daily and daily results from the
antenna testing are at the sub-millimeter level, based on the fact that 24-hour solutions
were used to average any possible bias. (3) A potential contributor that might have an
impact on the TAMDEF stations positioning is the pseudorange multipath effect; thus,
the root mean squared variations were estimated and analyzed in order to identify the
most and least affected sites. MCM4 was found to be the site with highest multipath, and
this is not good at all, since MCM4 is the primary ITRF access point for this part of
Antarctica. Additionally, results from the pseudorange multipath can be used for further
data cleaning to improve positioning results. (4) The Ocean Tide Modeling relied on the
use of two models: CATS02.01 (Circum Antarctic Tidal Simulation) and TPXO6.2
(TOPEX/Poseidon) to investigate which model suits the Antarctic conditions best and its
effect on the vertical coordinate component at the TAMDEF sites. (5) The scatter for the
time-series results of the coordinate components for the TAMDEF sites are smaller when
processed with respect to the Antarctic tectonic plate (Case I), in comparison with the
other tectonic plates outside Antarctica (Case II-IV). Also, the seasonal effect due to the
time-series seen in the TAMDEF sites with longer data span are site dependent; thus, data
processing is not the reason for these effects. (6) Furthermore, the results coming from a
homogeneous global network with coordinates referred and transformed to the ITRF2000
at epoch 2005.5 reflect the quality of the solution, obtained when processing TAMDEF
network data with respect to the Antarctic tectonic plate. (7) An optimal data reduction
strategy was developed, based on three different troposphere models and mapping
functions, tested and used to estimate the total wet zenith delay (TWZD) which later was
transformed to precipitable water vapor (PWV). PWV was estimated from GPS
measurements and validated with a numerical weather model, AMPS (Antarctic
Mesoscale Prediction System) and radiosonde PWV. Additionally, to validate the
TWZD estimates at the MCM4 site before their conversion into the GPS PWV, these
estimates were directly compared to TWZD computed by the CDDIS (Crustal Dynamics
Data Information System) analysis center. (8) The results from the Least-Squares
ii
adjustment with Stochastic Constraints (SCLESS) as performed with PAGES are very
comparable (mm-level) to those obtained from the alternative adjustment approaches:
MINOLESS (Minimum-Norm Least-Squares adjustment); Partial-MINOLESS (Partial
Minimum-Norm Least-Squares adjustment), and BLIMPBE (Best Linear Minimum
Partial-Bias Estimation). Based on the applied network adjustment models within the
Antarctic tectonic plate (Case I), it can be demonstrated that the GPS data used are clean
of bias after proper care has been taken of ionosphere, troposphere, multipath, and some
other sources that affect GPS positioning.
Overall, it can be concluded that no suspected of bias was present in the obtained
results, thus, GPS is indeed capable of capturing the signal which can be used for further
geophysical interpretation within Antarctica.
This report was prepared by Guadalupe Esteban Vázquez Becerra, a graduate student, Department of Civil and Environmental Engineering and Geodetic Science, under the supervision of Professor Dorota A. Grejner Brzezinska and Burkhard Schaffrin.; This research was supported by the National Council for Science and Technology (CONACYT) and partially supported by a grant from the National Science Foundation.; This report was also submitted to the Graduate School of The Ohio State University as a thesis in partial fulfillment of the requirements for the degree Doctor of Philosophy.
2009-07-01T00:00:00ZVázquez Becerra, Guadalupe Esteban, 1973-Ellipsoidal Wavelet Representation of the Gravity Field
http://hdl.handle.net/1811/78630
Ellipsoidal Wavelet Representation of the Gravity Field
Schmidt, Michael; Fabert, Oliver
The determination and the representation of the gravity field of the Earth are some of the most important
topics of physical geodesy. Traditionally in satellite gravity recovery problems the global gravity
field of the Earth is modeled as a series expansion in terms of spherical harmonics. Since the Earth’s
gravity field shows heterogeneous structures over the globe, a multi-resolution representation is an
appropriate candidate for an alternative spatial modeling. In the last years several approaches were
pursued to generate a multi-resolution representation of the geopotential by means of spherical base
functions.
Spherical harmonics are mostly used in global geodetic applications, because they are simple and the
surface of Earth is nearly a sphere. However, an ellipsoid of rotation, i.e., a spheroid, means a better
approximation of the Earth’s shape. Consequently, ellipsoidal harmonics are more appropriate than
spherical harmonics to model the gravity field of the Earth. However, the computation of the coefficients
of a series expansion for the geopotential in terms of both, spherical or ellipsoidal harmonics,
requires preferably homogeneous distributed global data sets.
Gravity field modeling in terms of spherical (radial) base functions has long been proposed as an
alternative to the classical spherical harmonic expansion and is nowadays successfully used in regional
or local applications. Applying scaling and wavelet functions as spherical base functions a
multi-resolution representation can be established. Scaling and wavelet functions are characterized
by the ability to localize both in the spatial and in the frequency domain. Thus, regional or even local
structures of the gravity field can be modeled by means of an appropriate wavelet expansion. To be
more specific, the application of the wavelet transform allows the decomposition of a given data set
into a certain number of frequency-dependent detail signals. As mentioned before the spheroid means
a better approximation of the Earth than a sphere. Consequently, we treat in this report the ellipsoidal
wavelet theory to model the Earth’s geopotential.
Modern satellite gravity missions such as the Gravity Recovery And Climate Experiment (GRACE)
allow the determination of spatio-temporal, i.e., four-dimensional gravity fields. This issue is of
great importance in the context of observing time-variable phenomena, especially for monitoring the
climate change. Global spatio-temporal gravity fields are usually computed for fixed time intervals
such as one month or ten days. In the last part of this report we outline regional spatio-temporal
ellipsoidal modeling. To be more specific, we represent the time-dependent part of our ellipsoidal
(spatial) wavelet model by series expansions in terms of one-dimensional B-spline functions. Thus,
our concept allows to establish a four-dimensional multi-resolution representation of the gravity field
by applying the tensor product technique
The research which led to this report was initiated during a visit to the Ohio State
University (OSU) from February 2002 through January 2003, and partially supported by
grants from the National Geospatial-Intelligence Agency's (NGA's) University Research
Initiative (NURI), entitled 'Application of spherical wavelets to the solution of the
terrestrial gravity field model' (NMA201-00-1-2006, 2000-2005, PIs: C. Shum), and from
the National Science Foundation's Collaboration in Mathematics and Geosciences (CMG)
Program, entitled 'Multi-resolution inversion of tectonically driven spatio-temporal
gravity signals using wavelets and satellite data' (EAR-0327633, 2003-2007, PIs: C.K.
Shum).
The authors benefited very much from discussions with many researchers at the OSU and
thank C.K. Shum and the OSU for hospitality. Further thanks go to the German Geodetic
Research Institute (DGFI) and the University of Munich (LMU), at which parts of the
work were conducted. Finally we thank Erik W. Grafarend for many fruitful discussions,
which were the actual starting point for this project.
2008-01-01T00:00:00ZSchmidt, MichaelFabert, OliverRadar Altimetry Methods for Solid Earth Geodynamics Studies
http://hdl.handle.net/1811/78629
Radar Altimetry Methods for Solid Earth Geodynamics Studies
Lee, Hyong Ki, 1975-
Satellite radar altimetry, which was initially designed for accurate measurements of sea
surface height, has been demonstrated to be applicable to non-ocean surfaces as well. In
this study, three different applications of satellite altimetry to geodynamics studies have
been examined: solid Earth crustal deformation due to Glacial Isostatic Adjustment (GIA)
over Hudson Bay, North America, water level fluctuation over vegetated wetlands of
Louisiana, and ice sheet elevation change over the Amundsen Sea sector, West Antarctica.
For efficient altimetry data processing, high-rate (10-Hz for TOPEX, 18-Hz for
Environmental Satellite (Envisat)) regional stackfiles were developed based on the
original low-rate (1-Hz) global ocean stackfile. A modified threshold retracker has also
been developed for more accurate land radar waveform retracking. 90-m resolution Cband
Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) plays
an important role to be used as a reference surface to select an optimal retracker, to
correct surface gradient errors, and to calculate land surface anomalies over Hudson Bay.
As a result, the crustal vertical motion is estimated from TOPEX decadal (1992-2002)
time series over land surfaces with height variation (in terms of standard deviation) less
than 2 m. The estimated vertical motion has been compared with the vertical motion
derived from Gravity Recovery and Climate Experiment (GRACE) satellite and several
GIA models. It agrees well with the laterally varying 3D GIA model, RF3S20 (β=0.4)
with differences of -0.1 ± 2.2 mm/year. It is anticipated that the new observation from
this study can be used to provide additional constraints for GIA model improvement. The
10-Hz stackfile procedure has also been utilized to observe the Louisiana wetland water
level variations over each of 10-Hz stackfile bin with along-track spacing of ~660 meter
using TOPEX altimeter measurements. The feasibility of applying retracking has also
been examined. Specifically, the water level variation over the swamp forest along the
Atchafalaya River basin has been examined with the SRTM DEM elevation and L-band
Advanced Land Observing Satellite (ALOS) Synthetic Aperture Radar (SAR) imagery. It
has been found that the water level fluctuations in terms of amplitude of each 10-Hz
TOPEX time series becomes larger as the elevation decreases and the SAR backscattered
power increases. Over the Amundsen Sea sector, which suffers dynamic thinning due to
the recent acceleration of coastal glaciers, the 18-Hz stackfile has been built using
Envisat retracked measurements. The rates of ice sheet elevation changes have been
estimated over 1° × 1° regions with 1-km resolution Antarctic DEM which is used to
correct for the surface gradient error. The ice mass loss from September 2002 – May
2005 has been estimated to be -49 ± 5 Gigaton/year after correcting for the firn depth
changes, which correspond to equivalent sea level change of 0.14 ± 0.01 mm/year.
This research is partially supported by grants from NOAA/NEDIS
NA06NES400007, NASA Earth Science program: NNX08AT52G, NNG04GN19G,
NNG04GN19G, JPL 1265252, 1283220, NSF's CMG Program (EAR0327633, and OCE-
0620885), Hydrology Program (EAR-044007), and by NGA NURI Program HM1582-
07-1-2024.
2008-09-01T00:00:00ZLee, Hyong Ki, 1975-Analysis of Stochastic Properties of GPS Observables
http://hdl.handle.net/1811/78628
Analysis of Stochastic Properties of GPS Observables
Vazquez Becerra, Guadalupe Esteban, 1973-
Traditionally, data processing for GPS positioning requires modeling
considerations underlying the observations. The variance-covariance (v-c) matrix (as
part of the stochastic model) usually comprises only the variances of the individual
pseudo-range and carrier phase observations and generally disregards any possible
correlation among them. However, for high precision, optimal GPS positioning
estimators, it might be important to account for the possible correlation between the GPS
observables.
The objective of this research is based on two fundamental considerations; the
primary one is related to the stochastic analysis of the different types of GPS observables
in order to estimate and interpret the level of the measurement noise (based on singledifference
residuals). For this purpose, a static survey (zero baseline) was performed
with six pairs of geodetic-grade GPS receivers of different type and make. Based on
these data, the normalized autocorrelation, cross-correlation, power spectral density
functions and histograms were thoroughly examined.
The secondary consideration is related to the construction of an alternative v-c
matrix, which implements the major outcomes of the stochastic analysis (auto and crosscorrelation
functions), in order to test its impact in the positioning estimators (coordinates
determination) using precise GPS positioning.
The results presented in this thesis showed that the different types of geodeticgrade
GPS receivers analyzed here possess distinct noise characteristics. In other words,
the noise characteristics are receiver specific. Furthermore, correlation exists among the
different types of GPS observables (cross-correlation) and it varies between the receivers.
In terms of positioning estimators, an example for zero and short baseline (10 m)
measurements was analyzed. In both cases (zero and short baselines), the results
obtained using the traditional approach (diagonal v-c matrix) better compare to “true”
values as opposed to those using the alternative v-c matrix, which accounts for
correlation among the observables. This indicates, that the results obtained in this case
study may not always apply to survey data, and more research is needed to formulate a
more generic model.
This report was prepared by Guadalupe Esteban Vazquez Becerra, a graduate student, Department of Civil and Environmental Engineering and Geodetic Science, under the supervision of Professor Dorota A. Grejner Brzezinska and Christopher Jekeli.; This Research was supported by the National Council for Science and Technology (CONACYT).; This report was also submitted to the Graduate School of The Ohio State University as a thesis in partial fulfillment of the requirements for the degree of Master of Science.
2008-10-01T00:00:00ZVazquez Becerra, Guadalupe Esteban, 1973-The Estimation Methods for an Integrated INS/GPS UXO Geolocation System
http://hdl.handle.net/1811/78627
The Estimation Methods for an Integrated INS/GPS UXO Geolocation System
Lee, Jong Ki
Unexploded ordnance (UXO) is the explosive weapons such as mines, bombs, bullets,
shells and grenades that failed to explode when they were employed. In North America,
especially in the US, the UXO is the result of weapon system testing and troop training
by the DOD. The traditional UXO detection method employs metal detectors which
measure distorted signals of local magnetic fields. Based on detected magnetic signals,
holes are dug to remove buried UXO. However, the detection and remediation of UXO
contaminated sites using the traditional methods are extremely inefficient in that it is
difficult to distinguish the buried UXO from the noise of geologic magnetic sources or
anthropic clutter items. The reliable discrimination performance of UXO detection
system depends on the employed sensor technology as well as on the data processing
methods that invert the collected data to infer the UXO. The detection systems require
very accurate positioning (or geolocation) of the detection units to detect and discriminate
the candidate UXO from the non-hazardous clutter, greater position and orientation
precision because the inversion of magnetic or EMI data relies on their precise relative
locations, orientation, and depth. The requirements of position accuracy for MEC
geolocation and characterization using typical state-of-the-art detection instrumentation
are classified according to levels of accuracy outlined in: the screening level with position
tolerance of 0.5 m (as standard deviation), area mapping (less than 0.05 m), and
characterize and discriminate level of accuracy (less than 0.02m).
The primary geolocation system is considered as a dual-frequency GPS integrated with a
three dimensional inertial measurement unit (IMU); INS/GPS system. Selecting the
appropriate estimation method has been the key problem to obtain highly precise
geolocation of INS/GPS system for the UXO detection performance in dynamic
environments. For this purpose, the Extended Kalman Filter (EKF) has been used as the
conventional algorithm for the optimal integration of INS/GPS system. However, the
newly introduced non-linear based filters can deal with the non-linear nature of the
positioning dynamics as well as the non-Gaussian statistics for the instrument errors, and
the non-linear based estimation methods (filtering/smoothing) have been developed and
proposed. Therefore, this study focused on the optimal estimation methods for the
highly precise geolocation of INS/GPS system using simulations and analyses of two
Laboratory tests (cart-based and handheld geolocation system).
First, the non-linear based filters (UKF and UKF) have been shown to yield superior
performance than the EKF in various specific simulation tests which are designed similar
to the UXO geolocation environment (highly dynamic and small area). The UKF yields
50% improvement in the position accuracy over the EKF particularly in the curved
sections (medium-grade IMUs case). The UKF also performed significantly better than
EKF and shows comparable improvement over the UKF when the IMU noise probability
iii
density function is symmetric and non-symmetric. Also, since the UXO detection
survey does not require the real-time operations, each of the developed filters was
modified to accommodate the standard Rauch-Tung-Striebel (RTS) smoothing algorithms.
The smoothing methods are applied to the typical UXO detection trajectory; the position
error was reduced significantly using a minimal number of control points. Finally, these
simulation tests confirmed that tactical-grade IMUs (e.g. HG1700 or HG1900) are
required to bridge gaps of high-accuracy ranging solution systems longer than 1 second.
Second, these result of the simulation tests were validated from the laboratory tests using
navigation-grade and medium-grade accuracy IMUs. To overcome inaccurate a priori
knowledge of process noise of the system, the adaptive filtering methods have been
applied to the EKF and UKF and they are called the AEKS and AUKS. The neural
network aided adaptive nonlinear filtering/smoothing methods (NN-EKS and NN-UKS)
which are augmented with RTS smoothing method were compared with the AEKS and
AUKS. Each neural network-aided, adaptive filter/smoother improved the position
accuracy in both straight and curved sections. The navigation grade IMU (H764G) can
achieve the area mapping level of accuracy when the gap of control points is about 8
seconds. The medium grade IMUs (HG1700 and HG1900) with NN-AUKS can
maintain less than 10cm under the same conditions as above. Also, the neural network
aiding can decrease the difference of position error between the straight and the curved
section. Third, in the previous simulation test, the UPF performed better than the other
filters. However since the UPF needs a large number of samples to represent the a
posteriori statistics in high-dimensional space, the RBPF can be used as an alternative to
avoid the inefficiency of particle filter. The RBPF is tailored to precise geolocation for
UXO detection using IMU/GPS system and yielded improved estimation results with a
small number of samples. The handheld geolocation system using HG1900 with a
nonlinear filter-based smoother can achieve the discrimination level of accuracy if the
update rate of control points is less than 0.5Hz and 1Hz for the sweep and swing
respectively. Also, the sweep operation is more preferred than the swing motion
because the position accuracy of the sweep test was better than that of the swing test.
This work was supported by a project funded by the US Army Corps of Engineers,
Strategic Environment Research and Development Program, contract number W912HQ-
08-C-0044.; This report was also submitted to the Graduate School of the Ohio State
University in partial fulfillment of the PhD degree in Geodetic Science.
2009-12-01T00:00:00ZLee, Jong KiThe Ohio State University Stackfiles for Satellite Radar Altimeter Data
http://hdl.handle.net/1811/78626
The Ohio State University Stackfiles for Satellite Radar Altimeter Data
Yi, Yuchan
This document describes the OSU stackfile database for satellite radar altimetry and
software that is used to access and maintain the database system. The stackfile database
system can be viewed as a reformatted version of Geophysical Data Record (GDR) data
products of satellite radar altimeters. A stackfile database is accessible using 2-
dimensional location indices of nominal ground tracks while the GDR products are
registered in time along actual ground tracks. The third dimension of a stackfile is the
repeat cycle of a satellite altimeter mission. The purpose of this document is to use it as a
user’s guide of the OSU stackfile databases installed on a unix/linux server.
The original version of stackfiles was designed by Gerhard L.H. Kruizinga, Center for
Space Research, University of Texas at Austin, 1994. Most part of the text in this
document was borrowed from the CSR technical memo "The New Stackfiles" originally
written by G.L.H. Kruizinga in the summer of 1994 and updated on August 21, 1998.
2010-05-01T00:00:00ZYi, YuchanCoseismic Deformation Detection and Quantification for Great Earthquakes Using Spaceborne Gravimetry
http://hdl.handle.net/1811/78624
Coseismic Deformation Detection and Quantification for Great Earthquakes Using Spaceborne Gravimetry
Wang, Lei
Because of Earth’s elasticity and its viscoelasticity, earthquakes induce mass
redistributions in the crust and upper mantle, and consequently change Earth’s external
gravitational field. Data from Gravity Recovery And Climate Experiment (GRACE)
spaceborne gravimetry mission is able to detect the permanent gravitational and its
gradient changes caused by great earthquakes, and provides an independent and thus
valuable data type for earthquake studies. This study uses a spatiospectral localization
analysis employing the Slepian basis functions and shows that the method is novel and
efficient to represent and analyze regional signals, and particularly suitable for extracting
coseismic deformation signals from GRACE. For the first time, this study uses the Monte
Carlo optimization method (Simulated Annealing) for geophysical inversion to quantify
earthquake faulting parameters using GRACE detected gravitational changes. GRACE
monthly gravity field solutions have been analyzed for recent great earthquakes. For the
2004 Mw 9.2 Sumatra-Andaman and 2005 Nias earthquakes (Mw 8.6), it is shown for the
first time that refined deformation signals are detectable by processing the GRACE data
in terms of the full gravitational gradient tensor. The GRACE-inferred gravitational
gradients agree well with coseismic model predictions. Due to the characteristics of
gradient measurements, which have enhanced high-frequency contents, the GRACE
observations provide a more clear delineation of the fault lines, locate significant slips,
and better define the extent of the coseismic deformation; For the 2010 Mw 8.8 Maule
(Chile) earthquake and the 2011 Mw 9.0 Tohoku-Oki earthquake, by inverting the
GRACE detected gravity change signals, it is demonstrated that, complimentary to
classic teleseismic records and geodetic measurements, the coseismic gravitational
change observed by spaceborne gravimetry can be used to quantify large scale
deformations induced by great earthquakes.
This Ohio State University Geodetic Science Report was prepared for, in part, and submitted to the Graduate School of the Ohio State University as a Dissertation in partial fulfillment of the requirements of the Doctor of Philosophy (PhD) degree.; This research is conducted under the supervision of Professor C.K. Shum, Division of Geodetic Science, School of Earth Sciences, The Ohio State University. The research results documented in this report resulted in a PhD Dissertation by Lei Wang (2012), Division of Geodetic Science, School of Earth Sciences, The Ohio State University. This research is partially funded by grants from NASA’s Interdisciplinary Science Program (NNG04GN19G), NASA’s Ocean Surface Topography Mission (OSTM) and Physical Oceanography Program (JPL1283230), the Air Force Materiel Command (FA8718-07-C-0021), and NSF’s Division of Earth Sciences (EAR-1013333). We would like to acknowledge Professor Frederik J. Simons, Department of Geosciences, Princeton University, for his hosting of Dr. Lei Wang for the summer visits.
2012-03-01T00:00:00ZWang, LeiDetection of a Local Mass Anomaly in the Shallow Subsurface by Applying a Matched Filter
http://hdl.handle.net/1811/78623
Detection of a Local Mass Anomaly in the Shallow Subsurface by Applying a Matched Filter
Abt, Tin Lian
The task is to locate a mass anomaly, particularly a void, in the near subsurface
based on gravity, gravity gradients, and magnetic field data. The motivation for
this search rises from multiple areas of applications such as urban planning, mining,
archeology, and extraterrestrial science. Assuming that the signal generated by the
sought mass anomaly is approximately known and can be described by the signal of a
three-dimensional prism, a Matched Filter (MF) is implemented to detect this signal
buried in the relatively strong noise of the geologic background. The background noise
is described in the filter function by covariances. One important aspect of the current
study is, therefore, to derive the covariance matrix that accounts for the relation
between gravity, gravity gradients, and the magnetic field. It turns out that the choice
of covariance function can have a significant influence on the MF performance. The
aim of this research is to answer some fundamental questions regarding the various
combinations of data types, the estimation of the depth or orientation of the mass
anomaly, as well as the optimal number of observed profiles. All tests are carried out
by Monte Carlo simulations, which include randomized simulated background fields.
In addition, a statistical interpretation based on the Neyman-Pearson hypothesis
test is provided. It determines the probabilities that either a sought anomaly is not
detected or that some background noise is mistaken for the sought anomaly. The
simulation results lead to the conclusion that the MF is a very strong tool to detect
the sought anomaly along an observed profile as it searches for the maximum Signal-to-Noise ratio. Therefore, the MF is able to detect anomalies buried in the background
noise even if they are not directly visible in the data set. A data set of either gravity
gradients or the magnetic field leads to more successful detections compared to a data
set of gravity. The combination of several gravity gradient components as well as the
magnetic field further improves the fileter performance. The MF is highly sensitive to
the depth of the sought anomaly but far less sensitive to its horizontal orientation.
As a consequence, it is possible to determine a rough estimate for the depth but
not for the orientation, which, however, can be approximated by measuring multiple
profiles.
This report was prepared for and submitted to the Graduate School of the Ohio State University
as a dissertation in partial fulfillment of the requirements for the PhD degree.
2011-08-01T00:00:00ZAbt, Tin LianA Comparison of Ellipsoidal and Spherical Harmonics for Gravitational Field Modeling of Non-Spherical Bodies
http://hdl.handle.net/1811/78622
A Comparison of Ellipsoidal and Spherical Harmonics for Gravitational Field Modeling of Non-Spherical Bodies
Hu, Xuanyu
Two harmonic expansions are compared on modeling the gravitational field of
non-spherical attracting bodies. As a solution to the Laplace’s equation based on
spherical coordinates, the spherical harmonic series (SHS) is uniformly convergent
outside a certain reference sphere that encloses the entire body mass. The
convergence can be doubtful in close proximity to the body. On the other hand, a
tri-axial ellipsoid, being more arbitrary-shaped than a sphere, is more apt to be
closely fitting to the attracting body. It would be desirable to apply the ellipsoidal
harmonic series (EHS) for field modeling if the body is distinctly non-spherical. To
obtain the EHS one solves the Laplace’s equation in ellipsoidal coordinates. In theory,
the EHS is akin to the SHS in terms of representation, properties and relation to the
field potential problem. It can be shown that EHS is convergent outside a certain
reference ellipsoid, thus could have more convergence region than the SHS.
However, the application of the EHS for field modeling is obscured by many
numerical difficulties. The theoretical formulation is far from practical and needs to
be modified at the cost of computational complexity. The numerical scheme of
applying the EHS for field modeling presented by Garmier and Barriot has been
reviewed and adopted for application in this work. The numerical accuracy of the
EHS may deteriorate from a certain degree, e.g., around 15, which limits the use of
higher-degree expansions.
In simulation we choose to compare the performance of EHS and SHS in
modeling the gravitational field of the Martian moon Phobos and asteroid 433 Eros,
both assumed to be homogeneous. Of the two bodies, Phobos has moderate shape
non-sphericity, while Eros is highly irregular-shaped. Results suggest that the EHS
and SHS models are comparable in performance in their respective convergence
regions. Outside the convergence region, both models are subject to divergence, i.e.,
incurring substantial modeling errors. And the further outside the convergence region,
the greater the errors. The divergence will be aggravated by increasing the degree of
expansions rather than be abated. With smaller convergence region than the EHS, the
SHS becomes more vulnerable to divergence in close range of the body, e.g., on the
reference ellipsoid. On the other hand, even if the EHS is applied outside its
convergence region, e.g., on the surface of the body, it is less error-prone than the
SHS, as the surface points usually reside at more shallow depth below the reference
ellipsoid than below the reference sphere. The comparison of simulation results on
Phobos and Eros suggest that the EHS is a more consistent model than the SHS for
the non-spherical bodies, at the expense of increased computational effort. And the
more non-spherical the body is, the more advantageous it is to apply the EHS.
This report was presented in partial fulfillment of the requirements for the degree
Master of Science in the Graduate School of The Ohio State University.
2012-06-01T00:00:00ZHu, XuanyuTerrain Corrections for Gravity Gradiometry
http://hdl.handle.net/1811/78621
Terrain Corrections for Gravity Gradiometry
Huang, Ou
This study developed a geostatistical method to determine the required extent of
terrain corrections for gravity gradients under the criterion of different applications. We
present the different methods to compute the terrain corrections for gravity gradients for
the case of ground and airborne gravity gradiometry. In order to verify our geostatistical
method and study the required extent for different types of terrain, we also developed a
method to simulate topography based on the covariance model. The required extents were
determined from the variance of truncation error for one point, or furthermore from the
variance of truncation error difference for a pair of points, and these variances were
verified with that from the deterministic method. The extent of terrain correction was
determined for ground gradiometry based on simulated, ultra-high resolution topography
for very local application, and also was determined based on mountainous topography of
large areas. For airborne gradiometry, we compute the terrain corrections and the
required extent based on Air-FTG observations at Vinton Dome, LA and Parkfield, CA
area; also they were verified with the results of Bell Geospace. Finally, from the mostly
flat, medium rough and mountainous areas, an empirical relationship was developed
which has the properties that the required extent has 4 times relationship corresponding to
the amplitude of PSD has 100 times relationship between mountainous and mostly flat
areas, and it can be interpolated for other types of topography from their geostatistics.
This report was prepared for and submitted to the Graduate School of the Ohio State University
as a dissertation in partial fulfillment of the requirements for the PhD degree.
2012-06-01T00:00:00ZHuang, OuOcean Tides Modeling using Satellite Altimetry
http://hdl.handle.net/1811/78620
Ocean Tides Modeling using Satellite Altimetry
Fok, Hok Sum, 1980-
Ocean tides, resulting from the gravitational attractions of the Moon and the Sun,
represent 80% of the ocean surface topography variability with a practical importance for
commerce and science over hundreds of years. Tides have strong influence on the
modeling of coastal or continental shelf circulations, play a significant role in climate due
to its complex interactions between ocean, atmosphere, and sea ice, dissipate their energy
in the ocean and solid Earth, and decelerate the Moon’s mean motion. Oceanographic
studies and applications, including coastal or continental shelf ocean circulations, also
require observations to be ‘de-tided’ using ocean tidal forward prediction models before
geophysical or oceanographic interpretation, particularly over coastal regions. Advances
in satellite radar altimetry technology enabled a globally sampled record of sea surface
height (SSH) and its changes over the past two decades, particularly after the launch of
TOPEX/POSEIDON satellite. This geophysical record enables numerous scientific
studies or discoveries, including improved global ocean tide modeling.
Several contemporary ocean tide models have been determined either through the
assimilation of satellite altimetry and coastal tide gauge data, often referred to as
‘assimilation models’ (e.g. FES2004, NAO.99b and TPXO6.2/7.1/7.2), or via the use of
altimetry observations in an ‘empirical modeling’ approach to solve for tidal constituents
based on a-priori tide models, including assimilated models (e.g. DTU10,
EOT08a/10a/11a, GOT00.2/4.7). However, ocean tide model accuracy is still much
worse, up to an order of magnitude, in the coastal regions or over partially or
permanently sea-ice or ice-shelf covered polar ocean, than that of models in the deep
ocean.
Here observation-based, empirical ocean tide models with 0.25°×0.25° spatial resolution,
the OSU12 models, has been determined using improved multi-satellite altimetry data
from TOPEX, Jason-1/-2, Envisat, and GFO, and based on a novel approach via spatiotemporal
combination, along with a robust estimation technique. We first demonstrate the
effectiveness of the spatio-temporal combination approach when comparing with various
ocean tide solutions under different data weighting schemes (i.e. equally weighted
solution, the weighted solution based on spatial (co-)variances, and the weighted solution
based on temporal (co-)variances). The generated tide models show substantial
improvement near coastal regions when compared against contemporary ocean tide
models using assessment from independent tidal constants of tide gauges and from
variance reduction studies using altimetry data. The improvement is particularly apparent
in regions with high hydrodynamic variability, yet the model accuracy is still regiondependent.
The model is available at: http://geodeticscience.org/oceantides/OSU12v1.0/
For the first time, the potential seasonality of ocean tides in subarctic regions has been
demonstrated. A statistically significant difference in variance reduction of multi-mission
altimeter SSH anomalies are observed in the subarctic ocean study region during summer
iii
and winter seasons. The variability in the SSH anomalies during winter are 15–30%
larger than those of summer, and we hypothesize that seasonality of tides contributes to
the observed SSH variability. The subsequent seasonal ocean tide solutions estimated
using observations only in the winter and in the summer seasons, reveal detectable
seasonal tidal patterns in the Chukchi Sea near the eastern Siberia region where it is
known to have seasonal presence of sea-ice covers.
This report was prepared for and submitted to the Graduate School of the Ohio State University as a dissertation for partial fulfillment of the requirements for the Doctor of Philosophy (PhD) degree.; This research is conducted under the supervision of Professor C.K. Shum, Division of Geodetic Science, School of Earth Sciences, The Ohio State University.; This research is partially supported by grants from NASA’s Physical Oceanography program, including Ocean Surface Topography Science Team Program (JPL 1356532, JPL 1384376, and UC154-5322), and from Ohio State University's Climate, Water, and Carbon (CWC) Program (http://cwc.osu.edu).
2012-12-01T00:00:00ZFok, Hok Sum, 1980-Topics in Total Least-Squares Adjustment within the Errors-In-Variables Model: Singular Cofactor Matrices and Prior Information
http://hdl.handle.net/1811/78619
Topics in Total Least-Squares Adjustment within the Errors-In-Variables Model: Singular Cofactor Matrices and Prior Information
Snow, Kyle Brian, 1962-
This dissertation is about total least-squares (TLS) adjustments within the errorsin-
variables (EIV) model. In particular, it deals with symmetric positive-(semi)definite
cofactor matrices that are otherwise quite arbitrary, including the case of crosscorrelation
between cofactor matrices for the observation vector and the coefficient
matrix and also the case of singular cofactor matrices. The former case has been
addressed already in a recent dissertation by Fang [2011], whereas the latter case
has not been treated until very recently in a presentation by Schaffrin et al. [2012b],
which was developed in conjunction with this dissertation. The second primary contribution
of this work is the introduction of prior information on the parameters to
the EIV model, thereby resulting in an errors-in-variables with random effects model
(EIV-REM) [Snow and Schaffrin, 2012]. The (total) least-squares predictor within
this model is herein called weighted total least-squares collocation (WTLSC), which
was introduced just a few years ago by Schaffrin [2009] as TLSC for the case of independent
and identically distributed (iid) data. Here the restriction of iid data is
removed.
The EIV models treated in this work are presented in detail, and thorough derivations
are given for various TLS estimators and predictors within these models. Algorithms
for their use are also presented. In order to demonstrate the usefulness of the
presented algorithms, basic geodetic problems in 2-D line-fitting and 2-D similarity
transformations are solved numerically. The new extensions to the EIV model presented
here will allow the model to be used by both researchers and practitioners to
solve a wider range of problems than was hitherto feasible.
In addition, the Gauss-Helmert model (GHM) is reviewed, including details showing
how to update the model properly during iteration in order to avoid certain pitfalls
pointed out by Pope [1972]. After this, some connections between the GHM and the
EIV model are explored.
Though the dissertation is written with a certain bent towards geodetic science,
it is hoped that the work will be of benefit to those researching and working in other
branches of applied science as well. Likewise, an important motivation of this work
is to highlight the classical EIV model, and its recent extensions, within the geodetic
science community, as it seems to have received little attention in this community
until a few years ago when Professor Burkhard Schaffrin began publishing papers on
the topic in both geodetic and applied mathematics publications.
This report is substantially the same as a dissertation that was prepared for and
submitted to the Graduate School of The Ohio State University for the PhD degree.
Except for the omission of some pages from the front matter, a different acknowledgment
page, and a change from double-space to single-space format, this report is
identical to the dissertation, which contains 15 pages with Roman numerals and 116
pages with Arabic numerals.
2012-12-01T00:00:00ZSnow, Kyle Brian, 1962-Applications of Synthetic Aperture Radar (SAR)/ SAR Interferometry (InSAR) for Monitoring of Wetland Water Level and Land Subsidence
http://hdl.handle.net/1811/78618
Applications of Synthetic Aperture Radar (SAR)/ SAR Interferometry (InSAR) for Monitoring of Wetland Water Level and Land Subsidence
Kim, Jin Woo, 1978-
Development of coastal wetlands and arid areas had negative impacts on the natural
hydrological processing on the surface and underground, and it resulted in disappearance
of wetlands that buffer severe flooding and function as home for various wildlife in the
wetlands, and groundwater depletion in the desert areas. Continuously monitoring the
surface change caused by human activities requires radar remote sensing with the in-situ
measurements. The intensity and phase components of Synthetic Aperture Radar (SAR)
data provide valuable information on the characteristics of surface change and ground
deformation. First of all, in this study, we demonstrated that the wetland water level
changes in the Atchafalaya Basin of the Louisiana can be effectively observed by
integrating Interferometric SAR (InSAR) results and radar altimetry data. When the
hydrologic flow between wetlands is disrupted by levees or dams, InSAR processing
cannot appropriately resolve the absolute water level changes from unwrapped phases.
The fusion of the two radar technologies enables one to accurately estimate absolute
water level change while avoiding inconsistent phase unwrapping. Secondly, the water
level in the Everglades is measured by monitoring stations, and the measurement is often
disturbed by abrupt water level rise. The L-band SAR backscatter coefficient in
Everglades has the characteristics that SAR intensity is inversely proportional with water
level in the freshwater marsh. The linear relationship enables one to estimate water level
from SAR backscattering coefficients. The correlation between two parameters over the
sawgrass was high, and it implied that water level estimation from the ALOS L-band
SAR backscatter coefficients is possible. The final study demonstrated the use of small
baseline subset (SBAS) InSAR processing technique to effectively measure the ground
subsidence caused by groundwater depletion in Tucson, Arizona. The SBAS processing
suppresses atmospheric artifacts affected by turbulent mixing that appears random in time
and space and estimates topographic error terms from multiple InSAR pairs. The SBAS
InSAR-derived vertical deformation gives information on the spatial extent and
magnitude of subsidence. The groundwater level decrease of tens of meter caused the
ground subsidence of tens of centimeters over a 17-years time period. InSAR results
indicate that the subsidence has recently slowed down possibly due to the artificial
recharge of water into surrounding aquifers near Tucson, Arizona.
This report was prepared for and submitted to the Graduate School of the Ohio State University as a dissertation in partial fulfillment of the requirements for the PhD degree.; This research is conducted under the supervision of Professor C.K. Shum, Division of Geodetic Science, School of Earth Sciences, The Ohio State University.; This research is primarily supported by NASA’s Earth and Space Science Fellowship (NESSF, No. NNX10AN54H), and by grants from National Geospatial-Intelligence Agency's (NGA’s) University Research Initiatives (NURI) Program (HM1582-07-1-2024 and HM1582-10- BAA-0002), the United States Geological Survey (G12AC20468), NASA Geodetic Imaging Program (NNX12AQ07G), and the Ohio State University's Climate, Water and Carbon program.
2013-08-01T00:00:00ZKim, Jin Woo, 1978-