Abstract for the COST 716 Workshop:
Towards Operational GPS Meteorology
Last updated: 02 July 2000
Please select author(s) to read the abstract
ATMOSPHERIC PROCESSING METHODS FOR
HIGH ACCURACY POSITIONING WITH THE GPS
Olivier BOCK (1) and Erik DOERFLINGER (2)
(1) Ecole Superieure des Geometres et Topographes, 72000 Le Mans, France
(2) Laboratoire de Geophysique, Tectonique et Sedimentologie, Universite de
Montpellier II, 30095 Montpellier cedex 05, France
The strategy for treating the tropospheric path delay in GPS data analysis
has been continuously evolving during the last decade, with the goal of
achieving a high accuracy in vertical positioning. The early approach,
which is still widely used, relies on the use atmospheric models and
surface meteorological data (either standard or measured). The main problem
with this approach is due to the small mixing of water vapour in the air,
hence remaining highly inhomogeneous and variable throughout the
atmosphere, which is thus difficult to model.
Two alternatives have emerged in the 90's, which can be referred to as
direct correction and estimation techniques. In the former techniques, the
effect of the tropospheric water vapour - the wet path delay - is inferred
from remote sensing instruments and GPS data are corrected prior to the
positioning process. Among these instruments, pointed water vapour
radiometers (WVRs) have been most widely experimented for GPS applications.
Other techniques, such as solar hygrometers and, more recently, IR
spectrometers and Raman lidars, have shown slightly better accuracy in
absolute water vapour sensing. Lidars have the main advantage of producing
profiles rather than integrated water vapour, and are thus likely to detect
small inhomogeneities. However, these more recent techniques need further
investigation for assessing their potential for achieving higher
positioning accuracy than WVRs currently do.
The estimation techniques are usually based on the standard procedure for
an a priori correction of the hydrostatic part of the path delay. The wet
path delay is, on the other hand, considered as a parameter which is
adjusted during the positioning process. It can be modelled either as a
series of (timely) independent deterministic parameters (least squares
approach) or as a stochastic variable (Kalman filtering). Recent
improvements of estimation models have been with the addition of horizontal
gradient parameters and the implementation of turbulent atmosphere models
(taking spatial correlations into account). Though the positioning accuracy
has only slightly been increased, these approaches merit further
investigation.
Direct correction and estimation techniques have shown comparable
positioning accuracy (commonly a few mm). Though the latter have the
advantage of requiring no specific site instrumentation, they might have
reached their ultimate accuracy in vertical positioning (due to the
mathematical limit in estimating correlated parameters). Hence,
instrumental improvements are still required for improving direct
correction techniques.
Another potential solution might be the use of data from 3D or 4D numerical
weather prediction models. This might have the advantage of requiring no a
priori mapping function, hence taking azimuthal asymmetry, both for the
hydrostatic and wet components, into account. But this requires first that
water vapour from various remote and in-situ sensors be properly
assimilated and that extensive humidity sensor networks be deployed. The
development of permanent GPS networks should largely contribute to this
latter point.
Johannes Böhm (1), Rüdiger Haas (2), Harald Schuh (1),
Robert Weber (1)
(1) Institute of Geodesy and Geophysics, Vienna University of Technology,
Austria
(2) Onsala Space Observatory, Chalmers University of Technology, Sweden
The estimation of geodetic parameters from GPS and VLBI requires the modeling of tropospheric refraction in order to account for the path delays of the radio signals being emitted from satellites and extragalactic radio sources respectively. Deviation from the azimuthal symmetry of the troposphere was described for example by Davis et al. (1993), introducing tropospheric gradients. For several years the estimation of tropospheric gradients has been included in the VLBI software package SOLVE (MacMillan, 1995). Recently, the VLBI software package OCCAM (modified Version 4.0) and the Bernese GPS software package (Version 4.2) allow gradient estimation, too. In this study we present comparisons of tropospheric gradients based on the analysis of VLBI and GPS data with the three software packages mentioned above. We concentrate on gradient estimates for Wettzell (Germany) derived from four VLBI experiments in early 1999 - two European sessions (Euro48, Euro49) and two intercontinental sessions (Iris-S 136, Iris-S 139) - and simultaneous GPS observations at Wettzell fundamental station.
Harald Bouma and Borys Stoew
Onsala Space Observatory Chalmers
University of Technology
Data from 56 GPS sites in Europe covering the period from 1997 until august 1999 provide information on the Total Zenith Delay (TZD). For seven of these information on the Integrated Water Vapor (IWV) is also available. Measurements are taken every hour. For every month the diurnal cycle in the TWD (or IWV) is estimated by averaging each data point referred to that particular hour for that month using the three years of available data. The more data we can use for averaging, the less influence we will have from specific weather events, such as the passage of a weather front, on the end result. To get a picture of the diurnal cycle of water vapor all over Europe, the IWV and the TZD are first compared at specific sites in order to determine how close relationship they have. The number of observations for each time period of the day should preferably be roughly the same. Preliminary results indicate low daily variations during the winter months and a larger daily signal during the summer months. We will also investigate possible dependence on the latitude of the site and the phase of the daily signal relative to the true solar time at each site.
Elmar Brockmann, Andreas Schlatter, Dieter Schneider, Thomas Signer and Adrian Wiget
Swiss Federal Office of Topography,
Seftigenstr. 264,
CH-3084 Wabern, Switzerland
In summer 1998, the Federal Office of Topography carried out two GPS / levelling campaigns to prove the capability of GPS for height estimation.
In the area of Emmental (peri-Alpine region with height differences of usually below 300 m) and in the area of Susten (Alpine region with height differences of up to 1600 m) GPS measurements were carried out on 4 consecutive days during both the day and the night. A morning and an afternoon session of 3 hours each was used for analyzing the height quality as a function of the session length. During the night sessions (8 hours) a larger network (baselines of 15-25 km) was observed to densify the GPS-based control network called LV95.
Alpine-scale height differences observed by GPS are biased mainly by a mismodelling of the troposphere. In collaboration with the ETH Zürich (Dr. A.Geiger), zenith path delays were computed from meteorological data of the Swiss Meteo Agency (SMA) using the 4-dimensional model called COMEDIE.
The campaigns were designed to answer the following topics:
- The impact of different processing strategies and different troposphere models (different a-priori models, estimation of tropospheric zenith path delays from GPS observations, introduction of zenith path delays from COMEDIE) on the GPS height estimation.
- Impact of data analysis using different GPS processing packages: the commercial packages GPSurvey 2.1 and SKI-Pro beta 1 and the scientific package Bernese 4.1
- Impact of the session length on the height estimates
- Accuracy of the GPS heights compared with levelling / geoid information
Elmar Brockmann, E. Calais, P. Pesec, F. Vespe, G. Weber
The five contries Austria, France, Germany, Italy and Switzerland which
are covering the alpine region as well as its surroundings agreed to equip
most of its permanent GPS-stations (all-together about 50 stations) with
one-hour download capabilities for near real-time zenith-delay estimation.
The paper describes the logistics of the network, the planned data-flow and
the distributed tasks of the contributing data- and analysis-centres. Whereas
in the first stage a pilot project of one month lenght is foreseen, it is
planned to continue the monitoring during the operational phase (WG-4) and
to merge the alpine station cluster and its products with similar clusters
to a future European meteorological GPS monitoring network.
J.M.Davila (1), J.Garate(1), M.Berrocoso(2)
(1) Real Instituto y Observatorio de la Armada. San Fernando. Cadiz (Spain)
(2) Depto. Matematica Aplicada. Universidad de Cadiz. (Spain)
The Royal Institute and Observatory of the Spanish Navy (Real
Instituto y Observatorio de la Armada en San Fernando, ROA) is located in
Cadiz, in the vicinities of the Gibraltar Strait, and is working on GPS
since mid 80's, mainly on geodetic-geodinamic and time-frecuency
applications. Since 1995, a permanent multipurpose GPS network is being
deployed in the south Spain-north Africa area, covering Andalucia, Gulf of
Cadiz, Alboran Sea and North Morocco. Nowadays, this net comprises four
permanent stations:SFER (San Fernando, SW Spain) an IGS station since the
end of 1995, CART (Cartagena, SE Spain), MELI (Melilla, North Africa) and
MAHO (Mahon, Balearic Is.). Both them are provided with TRIMBLE 4000 ssi or
sse receivers, Choke-ring or geodetic antennas, associated computer and are
controlled via phone line. MELI and MAHO stations have been installed under
MAGIG EU Project. Very Broad Band seismic stations (VBB) have been
co-instaled together with the GPS ones. Additional GPS stations focused on
time-frecuency (Turbo-Rogue receivers) are installed at the observatory
headquarters in San Fernando. Next step will be the densification of the GPS
network by the instalation of additional stations in Granada (South Spain)
and Alboran Island (mid Alboran Sea). In this work we will present the
present status of the net, the next future steps and the ROA GPS activities.
J.L. Davis
Harvard-Smithsonian Center for Astrophysics
Separating signals from noise is often a difficult task. In the case of ground-based GPS studies of the atmosphere, this task is made even more formidable by the fact that the atmospheric signal is itself noise, in the sense that it is a stochastic process. Nevertheless, this random nature can be exploited to develop different approaches to this problem. For example, networks of GPS sites may be used to separate signals that are common from signals unique to each site. We will present approaches for separating signal and noise, and discuss strengths and weaknesses of different approaches.
Michel Desbois, Remy Roca
Laboratoire de Meteorologie Dynamique, Ecole Polytechnique-CNRS, France
Water vapour is the major greenhouse gas in the atmosphere, keeping the
temperature of the Earth surface above the freezing level. It is also
highly variable in space and time, dependent on surface and air
temperature, and has a short life cycle in the atmosphere, of the order of
some days. It plays a large role in all the climate change processes,
firstly by its radiative feedback on the greenhouse change due to increased
CO2. Its transports and phase changes play also a large role in cloud and
precipitation processes, which in turn may influence the climate. These
processes are particularly active in the tropical regions, where they are
interacting with the vertical circulation cells. However, water vapour
integrated content and distribution are still not well captured as well
from conventional as from satellite measurements, and they are
insufficiently well simulated in climate models to allow proper predictions
of climate change. Although some observed regional trends in the total
water vapour content appear consistent in the last 20 years, there is still
no way to detect them from satellite radiometry, which requires accurate
calibrations. For the vertical distribution, which is very important for
the radiative effects, progresses are made from IR interferometric
instruments or microwave sounders, but the vertical profiles have to be
constrained by the total water vapour, as the low layers content is
difficult to retrieve from radiative measurements. It appears that GPS
techniques could be usefull i) for following the climatological trends and
variability of integrated water content at different time scales, ii) for
calibrating the satellites measuring the integrated water content, iii) for
providing directly or indirectly the constraint on total water vapour
necessary to inverse the water vapour profiles from satellite sounders.
Galina Dick, Gerd Gendt and Christoph Reigber
GeoForschungsZentrum Potsdam (GFZ),
Division Kinematics & Dynamics of the Earth,
Telegrafenberg, 14473 Potsdam, Germany
GFZ has started together with other research centers of the German
Helmholtz Society (HGF) the "GPS Atmosphere Sounding" Project (GASP)
on using ground-based (Subproject 1) and space-based (Subproject 2)
GPS observations for applications in numerical weather prediction
and climate research. First experience in Subproject 1 with
operational precision determination of water vapor in near real-time
within local GPS networks has demonstrated an accuracy of better
than 2 mm with a standard deviation at the level of better
than +- 1 mm. Densification of the test network is progressing
from 10 to 40 sites, requiring new analysis strategies. First
results from the enlarged network will be presented.
Jan Dousa
Geodetic Observatory Pecny
Research Institute of Geodesy,
Topography and Cartography
Zdiby 98
CZ - 250 66
It was well proved the accuracy together with the stability of
Near Real-Time (NRT) estimation of Precipitable Water Vapor (PWV)
from the network of ground-based GPS receivers principally depends
on orbit prediction. Good results in NRT schedule were achieved
if the information/accuracy-code of the individual orbit
quality was used in processing and selected parameters of orbits
were estimated simoultaneously. Therefore sub-daily orbit products
were requested allowing to minimize the prediction period.
We have tested the impact of new ultra-rapid orbits (produced twice
a day) on our routine PWV NRT estimation.
Gunnar Elgered
Onsala Space Observatory
Chalmers University of Technology
S-43992 Onsala, Sweden
Up to now GPS measurements of atmospheric water vapour have been used, as
an independent observational data set, to validate results from Numerical
Weather Prediction (NWP) models applied to forecasting activities and
climate research. Within the COST action we aim at the operational
applications as well as identifying the needs for requirements on a high
quality GPS data set for climate studies.
The primary objective is the assessment of the operational potential on
an international scale of the exploitation of a ground-based GPS system to
provide near real time observations for numerical weather prediction and
climate applications.
The COST action had its first Management Committee was held in Brussels
on January 8, 1999. The following 14 countries participate in the action:
Austria, Belgium, Denmark, Finland, France, Germany, Hungary, Italy,
Netherlands, Norway, Spain, Sweden, Switzerland, and United Kingdom.
The action is divided into four projects, each run by a working group:
(1) State of the art and product requirement, all participating countries
have reviewed their status and a final document is planned to be
distributed during the summer of 2000.
(2) A demonstration project, where a near real-time campaign will be
carried out early 2001, including all the steps from data acquisition to
assimilation into an NWP model.
(3) Applications, which includes the specifications from the
meteorology and climate communities. Here we stress the importance of
recognizing the different requirements of the GPS results depending on
the usage of the data.
(4) Planning for the operational phase. This project has not yet
commenced.
Ragne Emardson and Yoaz Bar-Sever
Jet Propulsion Laboratory, California Institute of Technology, California, USA
Jim Liljegren
Environmental Research Division, Argonne National Laboratory, Argonne,
Illinois, USA
We use a setup of a GPS receiver and a pointed water vapor radiometer
at the Cloud and Radiation Testbed (CART) site near Lamont, Oklahoma,
to asses the method of the extracting line-of-sight (LOS) tropospheric
delay from GPS. The data used in this study were collected over a
two-month period (Jul-Aug) 1999. During this time the GPS receiver
was able to track 12 satellites simultaneously. The WVR was programmed to
follow a schedule where it measured the atmospheric emission in the
direction of the GPS satellites
We processed the GPS data using the GIPSY software solving for a zenith
tropospheric delay and horizontal asymmetry. The LOS measurements were
then constructed from the zenith and gradient estimates as well as
the residuals. Comparisons with the direct WVR measurement shows various
degrees of agreement. We identify periods when the agreement is good and
when the results from the techniques differ significantly. We also discuss
different approaches in the processing of the GPS and WVR data. Finally we
address the benefits of LOS measurements in different areas of research.
Faccani C. (1), Ferretti R. (1), Ferraro C. (2),Nardi A. (2), Pacione R. (2), Sciarretta C. (2), Vespe F. (3)
(1) Dipartimento di Fisica, Universita' de L'Aquila
(2) Telespazio s.p.a. - Centro di Geodesia Spaziale Italiana-75100 Matera
(3) Agenzia Spaziale Italiana - Centro di Geodesia Spaziale Italiana-75100 Matera
A sub-set of European GPS network, equipped with surface
meteorological sensors, will be used to evaluate critically
the impact of GPS PW estimates on the weather prediction
by assimilating them into MM5 with the Nudging tecnique.
The selected network will be centered on Italy where, at present,
only few GPS stations are able to provide PW values.
In the near future this dense national GPS network
could be properly equipped if the GPS PW assimilation
turns out to be effective.
The PW maps obtained under various assimilation assumptions
(e.g. by including or exluding GPS PW and radiosoundings)
will be intercompared, trying to evaluate also the impact
of GPS PW values obtained with different processing strategies.
The results, properly handled, will also be compared to
directly measured quantities such as the water precipitation
recorded by rain-gauges.
A. Flores(1), A. Rius(2), J. Vila(3), A. Escudero(2)
(1) Institut d'Estudis Espacials de Catalunya, Barcelona, Spain.
(2) Institut d'Estudis Espacials de Catalunya, Barcelona, Spain.
(3) Institut d'Estudis Espacials de Catalunya, Barcelona, Spain.
We here present a tomographic reconstruction technique of the 4D structure of the wet refractivity in the lower troposphere using dense ground-based network of GPS receivers, with baselines up to 5 km. We have analyzed data from a campaign conducted at the Onsala Space Observatory during August 1998. Tomographic results have been validated using radiosonde data, the numerical weather model NCAR Mesoscale Model (MM5), satellite data from NOAA and analysis from the ECMWF. Results demonstrate that GPS tomography can provide accurate information in complex meteorological situations when mesoscale phenomena interact with local phenomena. Furthermore, atmospheric tomography using GPS data can provide a continuous spatio-temporal observation of water vapor content over the area of interest. It is therefore a potential valuable tool for atmospheric analysis, particularly interesting as an aid in situations and areas where models or punctual measurements fail to depict meteorological conditions.
M. Ge(1), E. Calais(1), J.Haase(2)
1) Centre National de la Recherche Scientifique, Géosciences-Azur, Sophia Antipolis, FR,
(2) ACRI Mécanique Appliquée et Sciences de l'Environnement, Sophia Antipolis, FR, BP. 234, F-06904 Sophia Antipolis,
We have been carrying out zenith tropospheric delay (ZTD) estimation from a network of about 50 permanent GPS stations spanning the Western Mediterranean since November 1998 in the framework of the MAGIC project. We will briefly review the network design, station distribution, GPS data flow, and GPS data processing strategy. We will present a new processing strategy for near real-time ZTD determination that allows the use of IGS predicted orbits with minimal degradation of the ZTD estimates. We will show the impact of the orbit estimation strategy, and the results from a comparison with IGS ultra-rapid orbits and Scripps hourly orbits. We will present results on the error sources associated with the ZTD estimate and establish criteria for error checking of the operational system.
Gerd Gendt, Galina Dick
GeoForschungsZentrum Potsdam (GFZ),
Division Kinematics & Dynamics of the Earth,
Telegrafenberg, 14473 Potsdam, Germany
Antonio Rius, Pepa Sedo
Institut d'Estudis Espacials de Catalunya,
CSIC Research Unit,
Edif. NEXUS 204, Gran Capita, 2-4,
08034 Barcelona, Spain
Within the GASP project ("GPS Atmosphere Sounding" Project) GFZ has
started activities at the end of 1998 using a small GPS ground-network at
synoptic sites of the German weather service. This network is operating
permanently since December 1998.
Using one selected week of data we made a comparison of two different
software packages and technologies developed at GFZ Potdam (EPOS.P.V2)
and IEEC Barcelona (based on GIPSY) for near real-time data analysis.
The comparisons comprise the NRT results as well as the post-processed
ones. The results are validated with water vapor radiometer measurements
and radiosondes data.
Lubomir Gradinarsky, Jan Johansson, Gunnar Elgered, and Per Jarlemark
Onsala Space Observatory
We present first results on the site co-ordinates and the atmospheric parameters from the permanent Global Positioning System site established at Chajnantor, in the Chilean Andes, at 5000 m altitude. The site was designed to ensure a long-term monument stability. The primary reason behind the site establishment was the future construction of the Atacama Large Millimeter Array (ALMA) at the same location. There is an interest to observe the site coordinate stability as well as the total radio path delay variations (the sum of the dry and the wet tropospheric components of the delay) which is possible employing the GPS technique. The fact that the site is known to have mainly a dry troposphere component makes it also a unique site for GPS accuracy studies. Included are also results from simulations of the impact of the orbit errors on the estimated atmospheric gradient parameters.
Seth I. Gutman(1), Tracy Lorraine Smith(2), Kirk L. Holub(3),
Michael Foy(3), Stanley G. Benjamin(1), and Barry Schwatz(1)
(1) NOAA Forecast Systems Laboratory, 325 Broadway, Boulder, Colorado 80303 USA,
(2) NOAA/FSL. Also affiliated with the Cooperative Institute for Research in the Atmosphere
(3) NOAA/FSL. Also affiliated with Systems Research Group, Inc.
Since 1994, the NOAA Forecast Systems Laboratory (FSL) has been
investigating the potential of ground-based GPS as a next-generation
upper-air observing system. A GPS Integrated Precipitable Water Vapor
(GPS-IPW) Demonstration Network has been established by FSL to
investigate the engineering and scientific bases of ground-based GPS
meteorology, and the impact that high temporal resolution IPW data have
on weather forecast accuracy. The network, currently consisting of 59
stations, is expected to grow to about 200 over the next few years in a
collaborative effort with several U.S. federal agencies and
universities.
We have determined that IPW can be measured using GPS and collocated
surface meteorological sensors with an error of less than 5% compared
with other observing systems. These measurements can be carried out
under all weather conditions with arbitrary temporal resolution. U.S.
numerical weather prediction (NWP) models typically have PW errors of
10%-15% under stable conditions (30%-50% under conditions of active
weather due primarily to the lack of timely and accurate moisture
observations), making accurate GPS-IPW a potentially significant proxy
quantity for NWP model assimilation. Since 1997, parallel NWP model
runs (with and without GPS, but each including all other available
surface and upper-air observations) have determined that the impact of
GPS-IPW on forecast accuracy is strongly influenced by the number of
stations
The accuracy of GPS zenith-scaled tropospheric signal delays from which
IPW is retrieved is strongly related to the accuracy of the GPS
satellite orbits. The use of 48-hour orbit predictions to calculate IPW
in near real-time (NRT) has resulted in encouraging but unreliable
results. This is primarily due to unanticipated spacecraft maneuvers
and difficulties in modeling the characteristics of a few satellites.
The latency with which GPS-IPW data is available depends on the
availability of improved GPS satellite orbits. This availability has
been reduced from about 2-weeks in 1994 (using precise IGS orbit
solutions), to about 36-hours in 1996 (using rapid orbits), to about
15-hours in 1999 due to the increased number of tracking stations
reporting daily and improved data processing. Since operational NWP
models are quickly moving to 1-hour forecast cycles, it is imperative
that improved GPS satellite orbits are available to support sub-hourly
data processing and model assimilation. Since late February, 2000, FSL
has been evaluating an experimental ultra-rapid GPS satellite orbit and
short-range (2-hour) prediction produced by Scripps Institution of
Oceanography. Initial results comparing NRT IPW computed with an 8-hour
sliding window for 30 days at 5 network sites and daily IPW solutions
indicate that sub-hourly data acquisition and processing is possible
with no significant bias and only a small (1.23 mm IPW) difference in
scatter. Comparisons of NRT GPS-IPW with radiosonde, and microwave
water vapor radiometer measurements will be discussed.
Jürgen Güldner, Deutscher Wetterdienst
Ground-based GPS for climate and NWP applications requires Integrated Water Vapour (IWV) data sets from independent instruments continuously available for quality control and accuracy estimation.
To meet the general need for high-quality water vapour information the WCRP/Global Water Vapour Project (GVaP) was initiated, which includes the establishment of reference observation stations. The Deutscher Wetterdienst is going to develop the Meteorologisches Observatorium Lindenberg (MOL) as a GVaP validation site.
First results of the intended automated monitoring and validation of IWV are shown using GPS-data, radiosondes, microwave profiler and microwave radiometer measurements.
J. Haase(1), H. Vedel(2), M. Ge(3), E. Calais(3)
(1) ACRI Mécanique Appliquée et Sciences de l'Environnement, Sophia Antipolis, FR, BP. 234, F-06904 Sophia Antipolis
(2) Danish Meteorological Institute, Copenhagen, DK
(3) Centre National de la Recherche Scientifique,Géosciences-Azur, Sophia Antipolis, FR
A year of GPS ZTD values for 44 sites in the Mediterreanean from the MAGIC project is compared to radiosonde data available in the same region. For most sites the difference has a positive bias (GPS ZTD greater than RS ZTD) and this bias increases with increasing humidity. For example, a higher humidity summer month at station ZIMM, has a bias of 16 mm ZTD as opposed to a winter month with 6 mm ZTD bias. The standard deviation also increases with increasing humidity. Coastal Mediterranean stations have much higher biases, (up to 21 mm), and a significantly different annual variability, in contrast to the smooth cyclic increase to higher ZTD in summer months typical of inland stations. This effect has not been observed in coastal stations outside the Mediterranean, such as in the United Kingdom. The differences between the GPS ZTD and RS ZTD are not correlated between sites. The time dependent bias has important implications for the assimilation of the GPS ZTD data, and may also have important implications for the vertical position of GPS sites used for sea-level studies in the Mediterranean.
J.Haase(1), E. Calais(2), J. Talaya(3), A. Rius(4), F. Vespe(5), R.
Santangelo(6), X.-Y. Huang(7), J.M. Davila(8), M. Ge(2), L. Cucurull(4), A.
Flores(4), C. Sciarretta(5), R. Pacione(5), M. Boccolari(6), S.
Pugnaghi(6), H. Vedel(7), K. Mogensen(7), X. Yang(7), J. Garate(8)
(1) ACRI Mécanique Appliquée et Sciences de l'Environnement, Sophia
Antipolis, FR, BP. 234, F-06904 Sophia Antipolis, jh@acri.fr;
(2) Centre National de la Recherche Scientifique,Géosciences-Azur, Sophia
Antipolis, FR
(3) Institut Cartogràfic de Catalunya, Barcelona, ES
(4) Institut d'Estudis Espacials de Catalunya, Barcelona, ES
(5) Agenzia Spatiale Italiana, Matera, IT
(6) Osservatorio Geofisico dell'Universita di Modena, Modena, IT
(7) Danish Meteorological Institute, Copenhagen, DK
(8) Real Instituto y Observatorio de la Armada, Cadiz, ES
The MAGIC project and the European COST Action 716 share common objectives for assessing the operational potential of ground-based GPS zenith tropospheric delay data in meteorology. The MAGIC project makes a significant contribution to all of the points in the COST Action definition: 1)development of a prototype ground based GPS system, 2)validation of the system, 3)development of data exploitation schemes for NWP and climate applications, and 4)requirements for operational implementation. The shared objectives have led to productive cooperation and motivated MAGIC participants to provide research results as part of the COST demonstration system. This poster gives an overview of the MAGIC project, highlighting the value of post-processed as well as near-real-time GPS ZTD data, and provides a guide to associated presentations that will be made by MAGIC project participants. MAGIC is a European Commission shared cost research project. http://www.acri.fr/magic
S. Hartig(1), E. Calais(1), J. Haase(2), F. Duquenne(3), M. Llubes(4), N.
Florsch(4), H. Vedel(5)
(1) CNRS, Géosciences-Azur, Sophia Antipolis, FR,
(2) ACRI Mécanique Appliquée et Sciences de l'Environnement, Sophia Antipolis, FR, BP. 234, F-06904 Sophia Antipolis,
(3) ESGT, 1, Rue Pythagore, 72000 Le Mans, FR
(4) CLDG, Université de La Rochelle, Avenue Marillac, 17042 La Rochelle, FR,
(5) Danish Meteorological Institute, Copenhagen, DK,
Under the leadership of the University of La Rochelle, and in collaboration with the ESGT (Le Mans), and the SHOM (Brest), we performed a one-week GPS experiment in Brittany at the end of October 1999, where the vertical motion due to ocean loading is expected to reach up to 15 cm at that time of the year. Seven GPS receivers were deployed in addition to the permanent sites of Brest and Le Mans. We investigated the impact of ocean loading effects on the estimation of zenith tropospheric delay (ZTD). We also used ZTD estimates from the HIRLAM numerical weather prediction model (provided by the Danish Meteorological Institute) as input to the GPS data analysis in order to improve the accuracy of the vertical component and better resolve the short-term vertical motion.
Mark Higgins,
The Met.Office, London Road Bracknell RG12 2ZS, UK
Mesoscale model analyses, produced using The Met.
Office's 3-dimensional variation analysis system, both with and without
total zenith
delay data will be presented and compared. We use delay data from a UK
network
and the CLIMAP network sites which fall within our mesoscale model
domain.
Jacob, D.
Max-Planck-Institute for Meteorology,
Bundesstr. 55, 20146 Hamburg, Germany
Water vapour plays a dominant role in the radiative balance and the hydrological
cycle. It is a principal element in the thermodynamics of the atmosphere, it
contributes absorbtion and emission in a number of bands and it condenses into
clouds that reflect and absorb solar radiation, thus directly affecting the
energy balance. Water vapour is the principal greenhouse gas in the atmosphere,
accounting for about 30 ° C warming due to longwave radiative trapping. In the
lower atmosphere, the water vapour concentrations can vary by orders of magnitude
from place to place. This variability creates a fundamental problem in climate
modelling. The contribution of water vapour for atmospheric phenomena on different
time and space scales for todays and future climates will be discussed as well as
the importance of water vapour monitoring. This is a prerequisite for model
validation and important help to understand atmospheric behaviour.
Per Jarlemark, Borys Stoew, Gunnar Elgered, Lubomir Gradinarsky, and Jan Johansson
Onsala Space Observatory
Chalmers University of Technology
S-43992 Onsala, Sweden
Satellite position and clock errors, antenna near-field
signal scattering and multipath are among the effects that
will contribute to the errors in the GPS observables used
in our regular 'Precise Point Positioning' solutions.
We have conducted Monte Carlo simulations in order to
quantify the influence of these effects individually.
Their influence on the atmospheric delay estimates at
different locations in a receiver network are discussed,
and characteristic correlation length scales of the errors
are presented.
Dave Jerrett
The Met. Office,
Beaufort Park, Easthampsted, Wokingham, Berkshire, RG40 3DN.
For a period of one week in August 1999 over sixty radiosonde ascents were made from the radiosonde station at Camborne SW England. Collocated at the site was a GPS receiver and Wind Profiler radar. During this period a number of occasions were observed in which the Integrated Water Vapour (IWV) calculated from the radiosondes departed consistently from that estimated by GPS. The detailed variation of IWV is compared to the synoptic situation and the Signal to Noise Ratios observed by the Wind Profiler. A possible explanation for the discrepancies between radiosondes and GPS IWV estimates is suggested.
D. Jerrett (1), O. Davies (2) and J. Nash (1)
1. The Met. Office, Beaufort Park, Wokingham, Berkshire,
RG40 3DN, United Kingdom
2. University Of Bath Electronic And Electrical Engineering,
Claverton Down
BA2 7AY, United Kingdom
The potential uses of GPS water vapour measurements for a variety of purposes, including use by local forecasters, numerical weather prediction and climate research will be reviewed.
Current user requirements for relative humidity will be interpreted in terms of equivalent total water vapour requirements.
Expected climatic changes in humidity will be quantified in terms of Total column water vapour over Europe. The Hadley Centre Climate model data is used to estimate the magnitude and distribution of IWV changes over Europe to be expected in the next 50 years.
The Total column water vapour amount is strongly correlated with in situ measurements that can be made at the surface. The largest impact from GPS IWV measurements will most likely come from those occasions when this correlation breaks down.
Some outstanding problems are highlighted, particularly the accuracy of
the Ocean tide loading correction and the impact of errors in the total
delay measurement in colder conditions. In terms of potential
operational GPS water vapour measurements for the UK it is suggested
that initial demonstration experiments concentrate on the problems of
forecasting summertime convection and fog in the southern UK. Successful
use of GPS for these purposes will depend on giving the forecaster the
ability to visualise the horizontal distribution of water vapour and to
be able to compare this with satellite products on a very short time
scale.
K.-P.Johnsen and B.Rockel
Institute of Atmospheric Physics
GKSS Research Center Geesthacht
In order to determine the quality of spaceborne and groundbased GPS data comparisons with the hydrostatic NWP model HRM of the german weather service and with data of the sensor DMSP-SSM/T2 will be presented. Between 1995 and 1998 GPS/MET aboard the MicroLab 1 successfully demonstrated the radio occultation technique to probe the Earth's atmosphere through their refraction effects on the signals transmitted by the GPS satellites. Profiles of the water vapor can be obtained from the GPS derived refractivity by using the temperature profiles of the model. The water vapor profiles and the ground-based measurements are used to compare with water vapor profiles and with vertically integrated atmospheric water vapor content, respectively, derived with the HRM and the SSM/T2.
F. Kleijer
University of Delft
Based on simulations and theoretic considerations several error sources in GPS-meteorology are studied. Special attention will be given to biases in air pressure, mapping functions and fixed coordinates of reference stations and elevation dependent weighting. These error sources are of special interest in water vapour estimation because they influence the estimates of the absolute tropospheric delay.
Antti Lange
Finnish Meteorological Institute P.O.Box 503, FIN-00101 Helsinki 10
It has been generally recognized that accurate horizontal gradients of Integrated Water Vapor (IWV) content of the atmosphere can be derived from the measurements of a regional network of geodetic GPS receivers. However, problems are encountered if only derivatives of observed values are available to data-assimilation in Numerical Weather Prediction (NWP). This problem can be overcome if an overall absolute level of the obtained IWV contents can be accurately estimated. The Statistical Calibration method proposed by Lange (1999) should make it possible to exploit meteorological and GPS data simultaneously for that purpose. The method is based on an optimal Gauss-Markov solution of a large regression equation system wherein the raw signal data from geodetic GPS receivers, the meteorological data from upper-air sounding stations and the departure data (innovation sequences) from HIRLAM runs are represented. It is expected that a suitable set of appropriate calibration parameters will be identified and that a fast Kalman filtering (FKF) method can be developed for operational estimation of them all as well as their error covariances. Such estimates are required for a truly optimal exploitation of the GPS measurements in Optimum Interpolation (O/I) and in 3- or 4-DVAR data-assimilation.
Alexander MacDonald, Yuanfu Xie
NOAA Forecast Systems Laboratory
Boulder, Colorado
Randolph Ware
University Corporation for Atmospheric Research & Radiometrics Corporation
Boulder, Colorado
Prediction of clouds and precipitation in the short range (0 to 36 hours) is economically very important, but is quite unskillful at the current state of the art of numerical weather prediction. There are many causes for this lack of skill. Perhaps the greatest deficiency is lack of knowledge about the initial state of the three dimensional water vapor field. Until recently, there was little hope of getting water vapor fields with the accuracy and detail needed for accurate mesoscale prediction. However, in the last few years new techniques have been developed for accurate measurement of slant water vapor (SW) -- the integrated water vapor along ray paths between GPS sensors and the GPS satellites in view. We present a technique based on three dimensional variational analysis (3DVAR) for the determination of the water vapor field using SW measurements from a network of surface based GPS sensors. We find that horizontal water vapor field structure is well determined by this technique, but vertical structue is not uniquely determined. By including water vapor soundings observed by microwave profilers in the simulated 3DVAR analysis (along with the GPS SW data) we find that unique three dimensional water vapor field structure can be obtained. These results have important implications for the improvement of short-range numerical weather prediction.
Beatriz Navascues (1), Jose Antonio Garcia-Moya (1), Lidia Cucurull (2), Antonio Rius (2)
(1) Instituto Nacional de Meteorologia, Madrid, Spain
(2) IEEC/CSIC Research Unit, Barcelona, Spain
The potential of GPS ground based data to improve the specification of the initial fields in NWP modelling is analysed. Real time GPS measurements at ground stations spread out over the Iberian Peninsula are compared to simulated values obtained from the Spanish Weather Service (INM) Limited Area Model (HIRLAM at INM) during a time period in spring 2000. Both analysis and short range forecasts model equivalents are generated. The forward operators of precipitable water (PW), and zenith total delay (ZTD) are presented, and shortcomings and advantages of the possible assimilated observed quantities. A comparison against close radiosondes stations data is carried out to know the error characteristics of this new observation system, including correlation of observation errors. This information will be essential when trying to assimilate this kind data by any analysis method.
A. E. Niell
MIT Haystack Observatory, Westford, MA 01886, USA
New mapping functions based on in situ meteorological parameters
have been developed for calculating the radio path length through
the atmosphere at elevations down to 3 degrees. The hydrostatic component
is based on the geopotential height of the 200 mb isobaric
pressure level. It provides a factor of two improvement in
accuracy and precision over previous hydrostatic mapping
functions at mid-latitudes. The wet component is calculated from
the profile of wet refractivity but will provide an improvement
of only about twenty-five percent. However, since the effect of
known errors in the hydrostatic mapping function dominates that
from the wet component, except near the equator, implementation
of these mapping functions should reduce the contribution of the
atmosphere to errors in estimates by VLBI and GPS of both the
vertical component of site position and the radio propagation
delay due to water vapor in the atmosphere.
T. Pany (1), P. Pesec (1), G. Stangl (2
(1) Austrian Academy of Science, Space Research Institute, Graz (Austria)
(2) Federal Office of Metrology and Surveying, Vienna (Austria)
Recent comparisons have shown that GPS inferred ZTD and ZTD calculated from numerical weather models already agree within 1-2 cm (rms). Therefore numerical weather models already describe the state of the troposphere quite well, such that these data set could be used to calculate and eliminate the influence of the troposphere on GPS measurements. Using the ECMWF global weather model we calculate the GPS path delays for each single GPS observation using a ray tracing procedure and subtract it from the GPS observations. We then process these corrected GPS data sets using undifferenced and double differenced observations and discuss the results. There is a number of problems like inter- and extrapolation in time and space in the numerical weather model, ray bending at low elevations, etc. which will be also addressed.
T.A. Springer
Since GPS week 1052, 5 March 2000, the IGS is producing a new combined orbit product. This product is called the IGS Ultra rapid orbit, IGU. The combined IGS Ultra rapid orbits are being made available twice every day, at 03:00 and 15:00 UTC, with a delay of 3 hours and are based on solutions from up to 7 different IGS Analysis Centers. Each ultra rapid orbit file covers 48 hours. The first 24 hours of the orbit are based on actual GPS observations (real orbit) the second 24 hours are extrapolated (predicted orbit). Like the IGS Predicted (IGP) orbits the Ultra rapid orbits are available for real-time usage. However, the quality of the Ultra rapid orbits should be significantly better because the average age of the predictions is reduced from 36 hours (IGP) to 9 hours (IGU). When the quality of the IGU products reaches a satisfactory level the IGU products will replace the IGP products. The main reason for the generation of the Ultra rapid products are the requirements, in both timeliness and accuracy, for near-real-time atmospheric monitoring, e.g., weather predictions. In this presentation we will show first results of the IGS Ultra rapid combinations. By comparing the IGU product to the other IGS products we will demonstrate the quality of this new product. In particular we will compare its quality to the IGP (predicted) product. Besides these comparisons we will also use the different IGS products to process a small GPS network and compare the results. The focus will be on the differences in the tropospheric zenith delay estimates because these are the parameters which are the "driving force" behind the IGS Ultra rapid products.
Peter Pesec
Space Research Institute, Austrian Academy of Sciences
As a basic ingredient for successfully covering the tasks of the COST-716
action "Exploitation of Ground-Based GPS for Climate and Numerical Weather
Prediction Applications" project 1 tried to find out the present status of the
technological status and development of the available GPS-monitoring stations.
The results are encouraging: By now about 200 European permanent GPS-stations
have the potential to contribute to hourly data dissemination, more than 10
analysis centres are able to provide hourly meteorological products. The final
report will give the required details which act as input for WG-2 and WG-3.
Borys Stoew, Jan Johansson and Gunnar Elgered
Onsala Space Observatory
Chalmers University of Technology,
439 92 Onsala, Sweden
The geodetic network of Global Positioning System (GPS) receivers generates data continuously with a temporal resolution of 30 s. These data are used to obtain total tropospheric delay estimates every 5 minutes. We present a recent development in the processing strategy which allows for the delivery of estimates of tropospheric parameters with small delay, based on the use of the predicted satellite orbits. The data are compared to the results from the processing that routinely delivers 7-day delayed estimates for the BALTEX project as well as to the results from a similar, near real-time activity within the MAGIC project.
Hans van der Marel
Delft University of Technology,
Department of Geodesy
Thijsseweg 11,
2629 JA Delft,
The Netherlands
In the framework of the European COST 716 action "Exploitation of Ground-Based GPS for Climate and Numerical Weather Prediction applications" a near real-time demonstration project is scheduled for 2001. The current plans foresee in two one month test datasets, in the summer and winter of this year, preceeding the more continuous demonstration starting in 2001. The two datasets in 2000 will be used to test the dataflow from GPS receiver, to GPS analysis center(s) and into the regular meteorological datastream for numerical weather prediction (NWP). The main objective of these tests is to make available two GPS datasets, in a format well suited for meteorology and archived like all other meteorological data, that can be used to study and further develop assimilation of GPS data into NWP models. These datasets are intented to assist in preparing the assimilation process. The assimilation itself will not be tested - in near real-time - until 2001, when the second phase of the demonstration project starts. The GPS processing is organised around several near real-time networks which contribute to this demonstration project. Some of these networks already exist, others are being set up at the moment. The networks are free to organise the processing as they seem fit in order to get, what they believe, the best possible results. The only requirements from the point of view of COST 716 are, that each GPS processing center computes properly validated Total Zenith Delays (TZD) with a well defined quality indicator, in one of the agreed formats. A significant amount of work will be needed in order to derive a good quality indicator and to understand the statistical properties of the TZD data. Various scenarios for the combination of data, or not, will be tested. However, in order to be of any use to NWP, the GPS derived zenith delays should be made available within 3 hours. More information can be obtained at the home page of the action (http://www.oso.chalmers.se/geo/cost716.html).
H. Vedel, K.S. Mogensen, Xiang-Yu Huang.
Danish Meteorological Institute,
Lyngbyvej 100,
DK-2100 Copenhagen, Denmark.
Beneficial use of gps-ztd's in numerical weather prediction and in
climate modeling and monitoring requires the ability to determine
precisely the ztd from meteorological data and that the ztd-measure is
useful to the models. We discuss various important aspects of the
calculation of ztd. We compare separately the dry and wet delays from
nwp-model hirlam and radiosonde reports from the MAGIC region
determining to which degree ztd is a pressure or a humidity measure
from a model point of view. This is of importance when making
assimilation algorithms for gps-ztd. As well as for the decission
whether to install pressure sensors at many gps-sites.
H. Vedel, K.S. Mogensen, Xiang-Yu Huang.
Danish Meteorological Institute,
Lyngbyvej 100,
DK-2100 Copenhagen, Denmark.
We present results of assimilation of gps derived ztd's. The ztd's are
from the MAGIC project common ztd database. The assimilation is done using a
prototype of the 3DVar program currently under development for HIRLAM.
The current knowledge about the error statistics of the gps ztd's in a
meteorological context is very limited, yet necessary for proper assimilation.
We analyse the importance of this problem by varying the error estimates and
the assimilation procedure.
A. Walpersdorf (1), E. Calais (2), J. Haase (3), L. Eymard (4) and M. Desbois (1)
(1) Laboratoire de Meteorologie Dynamique, Ecole Polytechnique, 91128 Palaiseau Cedex, France
(2) CNRS - Geosciences Azur, 250 Rue Albert Einstein, 06560 Valbonne, France
(3) ACRI, 260, Route du Pin Montard, 06904 Sophia Antipolis, France
(4) CETP, 10-12 Av. de l'Europe, 78140 Velizy, France
The estimation of horizontal atmospheric gradients in addition to
zenith delays is a strategy now commonly used in geodetic GPS
positionning, to compensate for varying atmospheric water vapor above
the GPS sites, and has shown to increase the measurement
precision. While the zenith delay has been successfully related to the
amount of integrated water vapor above the GPS site, the correlation
of the GPS estimated horizontal gradients with atmospheric quantities
remains unclear. To get a better understanding of the nature of these
gradients inferred by GPS, this study compares GPS observations from
the MAGIC permanent network around the Mediterranean Sea with
simulations based on the high resolution NWP model ALADIN (Meteo
France).
J. Wickert, R. Galas, R. König, Ch. Reigber
GeoForschungsZentrum Potsdam, Division 1
GPS ground station measurements are a basic input for the CHAMP Radio Occultation experiment. The quality and the way of providing these data affect in a crucial manner the results of atmospheric profiling and the potential use of its data products for the operational weather forecast. This presentation gives an overview of the CHAMP Radio Occultation experiment and will focus on the aspects of GPS ground network data use in occultation processing, in particular the schedule for providing the atmospheric data products in time for weather forecasting is discussed.