G51D-01 INVITED
GPS for estimation of TEC, electron density, velocity and scintillation in the ionosphere
Dual Frequency GPS time dealy and carrier phase observations provide a wealth of information about the ionosphere. The differential phase and time delay inform about TEC and the phase and amplitude of signals about the irregularities. TEC observations from multiple receivers can be used in a tomographic algorithm to produce maps of the spatial field of electron density. Such an approach can yield the ionospheric motion that is of particular interest during disturbed periods and at all times in high latitude regions where convection plays an important role in the ionospheric morphology. High frequency sampling of phase and amplitude at L1 can reveal scintillation events. Cross-correlation of the TEC, phase or amplitude from closely-spaced receivers reveals plasma irregularity velocity. Examples are drawn from the high-latitude ionosphere over the polar cap and Nother Europe. The discussion focusses on the existing receiver networks and improvements that could be made to the receiver distribution to optimise the ionospheric information that can be obtained.
G51D-02 INVITED
Ionospheric Challenges for GNSS Based Augmentation Systems
The ionosphere is a highly dynamic physical phenomenon that presents a variable source of error for Global Navigation Satellite System (GNSS) signals and GNSS based operational systems. The Federal Aviation Administration's (FAA) Wide-Area Augmentation System (WAAS) was designed to enhance the GNSS standard positioning service by providing additional accuracy, availability and integrity that is sufficient for use in commercial aviation. It is the first of a number of planned regional Satellite Based Augmentation Systems (SBAS). Other systems in development include the European EGNOS system, the MSAS system in Japan and the GAGAN system in India. In addition, the South American countries are investigating the feasibility of operating an SBAS system in this region. Much of the WAAS ionospheric research and development focused on defining and mitigating ionospheric challenges characteristic of the mid-latitude regions, where the ionosphere is well studied and relatively quiescent. The EGNOS and MSAS systems will primarily operate under a similarly quiescent mid-latitude ionosphere. SBAS system development in South America, India and other low-latitude regions, however, will have to contend with much more extreme conditions. These conditions include strong spatial and temporal gradients, plasma depletions and scintillation. All of these conditions have a potential to limit SBAS performance in the low latitude regions. This presentation will review the effects that the ionosphere has on the mid-latitude WAAS system. It will present the techniques that are used to mitigate ionospheric disturbances induced on the system during severe geomagnetic activity and it will quantify the effect that this activity has on system performance. The presentation will then present data from the South American Low-latitude Ionospheric Sensor Network (LISN) that can be used to infer the ionospheric effects on SBAS performance in the most challenging low-latitude ionospheric environment. LISN is a network of GNSS receivers and other ionospheric sensors that are in the process of deployment and installation across the western half of South America. Its purpose is to address key questions about the physics of the equatorial ionosphere and to develop forecasting/predictions capabilities concerning the onset of equatorial Spread F. The International Civil Aviation Organization has committed to transition to satellite navigation. This presentation will summarize the effects and the limitations that the ionosphere places on satellite based navigation systems.
G51D-03 INVITED
Mitigation of Ionospheric Effects on DGPS and WADGPS Operations
Under high levels of ionospheric activity, significant degradations in differential GPS (DGPS) and wide area DGPS (WADGPS) positioning accuracies can occur. DGPS and WADGPS methods are employed for many applications and millions of users. Examples include marine DGPS services, land applications (such as transportation monitoring, fleet management and emergency response) and commercial aviation. In previous studies for the North American sector, DGPS and WADGPS positioning errors were observed to increase by factors of 10-30 under increased ionospheric activity. In particular, gradients of up to 50 ppm are associated with a feature known as storm enhanced density (SED). This feature is a localized enhancement of total electron content (TEC) extending northwest through the mid-latitudes. Positioning errors of 20 m or more have persisted for hours during such events. Specific WADGPS services include the Satellite-Based Augmentation Systems (SBAS) WAAS and EGNOS. The WAAS has been designed for commercial aviation in the United States; EGNOS operates in a similar manner for the European sector. In this presentation, DGPS, WAAS and EGNOS capabilities are assessed under severe ionosphere events. The horizontal and vertical positioning accuracies are determined throughout North America and Europe during such events using available data from existing GPS networks (IGS and CORS). All DGPS baselines of length 100-200 km are processed; one station of the baseline is designated as reference and the other as remote user. Differential corrections are computed for the reference and applied at the remote user location. DGPS positioning solutions are generated for hundreds of baselines simultaneously in this manner to derive full spatial statistics of positioning accuracies. WADGPS positioning solutions are also generated for the same set of remote user stations using archived WAAS and EGNOS messages, and computing and applying localized corrections. An extensive study is conducted using data representative of a range of ionospheric conditions, including three major geomagnetic storm events. Under enhanced ionospheric activity, the WADGPS positioning errors can be as large as 25-30 m, and typically exceed those for single baseline DGPS. Methods of mitigating the larger positioning errors are proposed, with specifications for reference network densities and update intervals required. Space weather warning products designed for navigation applications are also introduced for the GNSS user community.
G51D-04
A Comprehensive Evaluation of the Errors Inherent in the use of a 2-Dimensional Shell for Modeling the Ionosphere
The ionosphere is a complex, dynamic part of the Earth's atmosphere and impacts many different frequencies of electromagnetic propagation passing through it, such as those of the Global Positioning System (GPS). For this reason, the ionosphere is frequently studied and modeled so as to remove these impacts. Unfortunately, because of the complexity of the ionosphere, such models frequently make use of a so-called "2 dimensional shell", for simplicity of both modeling and data distribution. In 2004, a joint study between NOAA's National Geodetic Survey (NGS) and Space Environment Center (SEC) led to the development of SEC's operational "USTEC" model– a truly four dimensional representation of the ionosphere over the conterminous United States produced in near-real time from GPS data collected at Continuously Operating Reference Stations (CORS). Because USTEC is spatially 3 dimensional, with values changing every 15 minutes (4-D), it allows for comprehensive examination of the complete distribution of electrons in the ionosphere over the U.S. Using USTEC, a study was performed to quantify the errors inherent in using a 2-D shell model, versus the more realistic 3-D model of the ionosphere. Errors of several percent were identified from purely geometrical sources related to the simplistic mapping function used in conjunction with 2-D shell models, but even larger errors were identified which were due to the inability of shell models to reflect even the most basic vertical mixing implicit in 3-D models. After quantifying all geometric errors associated with the well-known "cosine mapping function", an entirly new mapping function has been proposed to replace the cosine mapping function. This new function also relates the slant TEC to the vertical TEC within a shell model, just like the cosine mapping function, but removes as much as 50% of the errors inherent in the cosine mapping function. The final recommendation of this study is that users who require 1 TECU or better accuracy (approximately one L1 wavelength in GNSS positioning) should generally use a 3-D model of the ionosphere. Users who must rely on a 2-D model (for computational or bandwidth limits) should consider using the new mapping function in order to avoid large systematic errors associated with the long standing cosine mapping function.
G51D-05 INVITED
Numerical Weather Models for Troposphere Delay Modelling at Radio Wavelengths
Data from numerical weather models (NWM) have been successfully used to develop and improve models for the tropospheric delays as applied in the analysis of Global Positioning System (GPS), Doppler Orbitography and Radiopositioning Integrated by Satellite (DORIS) or Very Long Baseline Interferometry (VLBI) observations. Tropospheric delays are usually divided into a hydrostatic and a wet part, and both of them are set up as the product of the respective zenith delay and mapping function. Whereas the hydrostatic zenith delay can be determined from empirical models, 6 hour surface pressure values from NWM or in situ pressure recordings, the wet zenith delay is estimated in the analysis of space geodetic data. The mapping functions which account for the elevation dependence of the delays can be derived routinely from 6 hour NWM data like the Vienna Mapping Functions (VMF1), or they can be represented as an empirical model like the Niell Mapping Functions (NMF) or the Global Mapping Functions (GMF), the latter being consistent with the VMF1 over longer time spans. State of the art analyses also include the estimation of tropospheric gradients which account for the azimuthal asymmetries of the delays. We compare the influence of different types of a priori gradients (empirical model, station dependent values, 6 hour values from NWM) on the terrestrial and celestial reference frame in GPS and VLBI analysis.
G51D-06
Ray-traced tropospheric total slant delays for GNSS processing
Numerical weather models have undergone an improvement of spatial and temporal resolution in the recent years, which made their use for GNSS applications feasible. Ray-tracing through such models permits the computation of total troposphere delays and ray-bending angles. At the National Institute of Information and Communications Technology (NICT), Japan the so-called KAshima RAy-tracing Tools (KARAT) have been developed which allow to obtain troposphere delay corrections in real-time. Together with fine-mesh weather models from the Japanese Meteorological Agency (JMA) huge parts of the East Asian region, including Japan, Korea, Taiwan and East China, can be covered. The Japanese GEONET with its more than 1300 GNSS receivers represent an ideal test-bed for the evaluation of the performance of KARAT. In cooperation with the Geographical Survey Institute (GSI), Japan more than 1.6 billion observations, covering measurements from July 1st until August 31st, 2006, were processed and the corresponding troposphere delays were used to modify the original RINEX files by subtraction of code- and phase delays. These modified observations were processed by a dedicated analysis run of the GEONET operation center, taking advantage of the computer cluster at GSI. First results from this study, together with an in-depth discussion about the assets and drawbacks of the reduction of troposphere total slant delays will be given in this presentation. Additionally an overview about KARAT, the treatment of observational data and the impact of future refined numerical weather models on GNSS analysis will be included in this contribution.
G51D-07
Tropospheric Signal Delay Estimates Derived from Numerical Weather Prediction Models and Their Impact on Real-Time GNSS Positioning Accuracy
The accurate characterization of atmospheric moisture fields (including water vapor and clouds) is essential for improved weather forecasting and climate monitoring. Despite its importance, the ability to do so under all weather conditions has been a continuing problem for atmospheric scientists. The principle reason why this problem has been so difficult to solve is related to the high temporal and spatial variability of water in the free atmosphere. Under certain circumstances the distribution of moisture in the atmosphere can change abruptly over short distances, and this causes it to be under-observed using conventional weather observing systems. As water vapor, temperature and pressure change in the atmosphere, the refractivity of the troposphere changes accordingly and GNSS accuracy can suffer if the hydrostatic and wet signal delays are mismodeled. Recognizing this, the geodetic community developed techniques to treat the signal delays caused by the neutral atmosphere as nuisance parameters and remove them for high accuracy positioning applications. In ground-based GNSS/GPS Meteorology at NOAA, the tropospheric signal delay is estimated in near real-time from a network of about 400 continuously operating reference stations distributed across the U.S. using an 8-hr sliding window technique. Estimates of tropospheric refractivity (and/or integrated precipitable water vapor retrieved from these delays) have been assimilated into numerical weather prediction models in the U.S., Canada, Europe and Japan with exceptionally good results. Based on these and other findings, GNSS/GPS-Met is scheduled to transition from NOAA Research into operational use in NOAA's National Weather Service starting in 2009. Recognizing the need for improved ways to mitigate tropospheric effects on GNSS accuracy, especially for applications requiring low-latency measurements of height, scientists at NOAA's Earth System Research Laboratory began to investigate the feasibility of using operational numerical weather prediction (NWP) models to estimate the wet and dry refractivity of the troposphere to assist in integer-cycle ambiguity resolution and improve vertical position accuracy. The NOAA Tropospheric Signal Delay Model (NOAATrop) uses the state variables derived from NOAA's Rapid Update Cycle NWP model that assimilated GPS along with all other available atmospheric measurements to provide real-time zenith hydrostatic and wet signal delays, and their horizontal gradients, with a 2D RMS error of < 2.5 cm in the cool season and < 5 cm in the warm season. In this presentation, we will take a look "under the hood" of the NOAATrop model, explain how it works, give some examples from real-time GPS networks, and present some possible applications to the Earth System Science community. http://gpsmet.noaa.gov
G51D-08
Error Evaluations of GPS Estimates in Active Weather Conditions Based On High-Resolution Numerical Weather Model
Mitigations of effects of microwave propagation delay due to water vapor have been discussed in GPS communities for more than 10 years. It is furthermore important in the current era of real-time monitoring of kinematic coordinates and precipitable water vapor (PWV). The significant errors in coordinate estimates are usually observed in summer and in active weather conditions, and they should be mitigated to prevent misleading in geodynamical discussions. Quantitative estimation of errors in GPS estimates in such cases and understandings of mechanism of the errors are important for further improvement of GPS tropospheric models. In the present circumstances, the output from numerical weather model are only 3-dimensional data which can be used to validate atmospheric models used in GPS processing if the model can reproduce realistic variations of water vapor. We thus compute errors of GPS estimates from high-resolution NWP output and discuss limitation of the current tropospheric models used in GPS processing. We focused on several active weather conditions with severe precipitation, and performed high-resolution simulations with Weather Research and Forecasting (WRF) model. To reproduce realistic variations of water vapor, the horizontal scale was set to 4 km to avoid a cumulus parameterization. The forecast output is every 30 seconds which is equal to sampling rate in general GPS observation. We computed slant wet delay by ray-tracing of wet refractivity with actual GPS orbit information at real (SUOMINET) and imaginary GPS stations in the U.S. Wet mapping function, tropospheric wet delay gradients, and zenith wet delays were computed based on the slant wet delay, where reasonable constrains which is used in GPS processing were applied in the estimates. The estimates based on whole coverage of sky were also computed to evaluate errors came from inhomogeneous coverage of GPS satellite in particular axis (usually north-south axis). The residuals of slant wet delays from model fits are used to estimate coordinate errors. We also estimated representive errors for time and space in GPS retrieved precipitable water vapor (PWV) due to limitation of GPS tropospheric models and limited coverage of GPS satellites in the sky. Such information would be useful to decide both observation and representive errors used in data assimilation. http://www.cosmic.ucar.edu/~iwabuchi