1: %
2: % AA vers. 4.01, LaTeX class for Astronomy & Astrophysics
3: %
4: %___________________________________________________________ Ionosphere paper
5: % Version Feb'99
6: %
7: %\documentclass[referee]{aa}
8: %
9: \documentclass{aa}
10: %
11:
12: %\def\verbfont{\small\tt}
13: \def\et{et al.}
14: \def\deg{$^\circ$}
15: \def\degp{$\rlap{.}^\circ$}
16:
17: %\addtolength{\topmargin}{-0.3cm} % Original text
18: %\addtolength{\topmargin}{0.5 cm} % Original text
19:
20: %\addtolength{\textheight}{1 cm}
21: %\renewcommand\topfraction{0.9}
22: %\renewcommand\dbltopfraction{0.8}
23: %\renewcommand\bottomfraction{0.5}
24:
25: \def\aap{A\&A}
26: \def\apj{ApJ}
27: \def\apjs{ApJS}
28: \def\aj{AJ}
29: \def\aaps{A\&AS}
30: \def\aas{A\&AS}
31: \def\aar{A\&AR}
32: \def\araa{ARA\&A}
33: \def\baas{BAAS}
34: \def\sci{Sci}
35: \def\mnras{MNRAS}
36: \def\rsci{Radio Science}
37: \def\jgr{J.\ Geophys.\ Res.}
38:
39: \begin{document}
40:
41: \thesaurus{11 % Astrometry
42: (02.16.1; % Plasmas
43: 03.01.2; % Atmospheric effects
44: 03.13.5; % Methods: observational
45: 03.20.2; % Techniques: interferometric
46: 05.01.1)} % Astrometry
47: %
48: \title{A GPS-based method to model the plasma effects in VLBI observations}
49:
50: \author{E.\ Ros\inst{1,2}
51: \and
52: J.M.\ Marcaide\inst{1}
53: \and
54: J.C.\ Guirado\inst{1}
55: \and
56: E.\ Sard\'on\thanks{%
57: {\it Present address:\/} Grupo de Mec\'anica del Vuelo S.A.,
58: Isaac Newton 11, PTM Tres Cantos, E-27860
59: Madrid, Spain}\inst{3}
60: \and
61: I.I.\ Shapiro\inst{4}
62: }
63:
64: \offprints{E.\ Ros, ros@mpifr-bonn.mpg.de}
65:
66: \institute{
67: Departament d{'}Astronomia i Astrof\'{\i}sica,
68: Universitat de Val\`encia, E-46100 Burjassot, Val\`encia, Spain
69: \and
70: Max-Planck-Institut f\"ur Radioastronomie,
71: Auf dem H\"ugel 69, D-53121
72: Bonn, Germany
73: \and
74: Deutsche Forschungsanstalt f\"ur Luft- und Raumfahrt e.V., Fernerkundungstation
75: Neustrelitz, D-17235 Neustrelitz, Germany
76: \and
77: Harvard-Smithsonian Center for Astrophysics, 60 Garden St.,
78: Cambridge, MA 02138, US
79: }
80:
81: \date{Received 17 September 1999 / Accepted 17 January 2000} %5 January 2000}
82: \titlerunning{GPS-method to model plasma effects in VLBI observations}
83:
84: \maketitle
85:
86: \begin{abstract}
87: Global Positioning System (GPS) satellites broadcast at frequencies
88: of 1,575.42\,MHz (L1) and of 1,227.60\,MHz (L2). The dispersive property of
89: the ionosphere can be used to combine independent measurements
90: at the two frequencies to estimate
91: the total electron content (TEC) between a GPS receiver
92: site and a broadcasting satellite. Such measurements, made at
93: sites near to Very Long Baseline Interferometry (VLBI)
94: sites, can be used to estimate the
95: ionospheric contribution to VLBI observables.
96: For our 1991.9 astrometric VLBI experiment in which we obtained
97: group-delay observations in the 8.4 and 2.3\,GHz bands simultaneously, we
98: found that the GPS and VLBI determinations of the ionosphere delays
99: agreed with root-mean-square differences below
100: 0.15\,ns for intercontinental baselines and 0.10\,ns for continental
101: ones. We also
102: successfully applied the GPS-based procedure
103: to reduce the ionospheric effect in phase delays used for high
104: precision differenced astrometry at 8.4\,GHz for this same experiment.
105: \end{abstract}
106:
107: \keywords{Plasmas -- Atmospheric effects --
108: Methods: observational -- Techniques: interferometric -- Astrometry}
109:
110: \section{Introduction}
111:
112: VLBI provides unprecedentedly accurate angular resolution
113: through observations of celestial bodies with
114: radio telescopes spread over the Earth's surface.
115: Each observing
116: station records data on magnetic tapes. The local-oscillator
117: signals and the time-tagging of the data are governed by hydrogen-maser
118: frequency standards. The tapes are processed
119: in special-purpose correlators to determine the so-called VLBI observables:
120: group and phase delays and phase-delay rates.
121:
122: A main problem in determining the sky positions of celestial radio sources
123: from these VLBI observables is the effect of the Earth's ionosphere on them.
124: The use of GPS satellite data to remove this effect forms the thrust of
125: this paper.
126:
127: The ionosphere is characterized by
128: its content of free electrons and ions.
129: The F$_2$ layer of the ionosphere has the largest density of charged particles,
130: with values up to $3\times10^{12}$\,m$^{-3}$.
131: %Our accuracy goals can be achieved by modeling
132: %the ionosphere as a thin shell F$_2$ layer at a height of 350\,km
133: %above the Earth's surface (Davies \& Hartmann \cite{dav97}).
134: The total electron content per square meter along a line of sight
135: is the number of electrons in a column of
136: one square meter cross section along the ray path:
137: \begin{equation}
138: {\rm TEC}= \int_0^{h_0} N\cdot dh,
139: \label{equ:tec-def}
140: \end{equation}
141: where $N$ is the spatial density of electrons,
142: $h$ is the coordinate of propagation of the wave, and $h_0$ corresponds
143: to the effective end of the ionosphere.
144: TEC is highly variable and depends on several factors,
145: such as local time, geographical location, season, and solar activity.
146: TEC can have values between 1 TECU (or TEC unit, defined as
147: 10$^{16}$\,m$^{-2}$) and 10$^3$ TECU. Epochs of greater solar
148: activity cause higher values of the TEC.
149:
150: The ionosphere affects the phase and group delays oppositely (to first
151: order, see, e.g.,
152: Thompson \et\ \cite{tho86}):
153: \begin{equation}
154: \Delta \tau = \mp \frac{\kappa}{c\nu^2}\cdot {\rm TEC},
155: \label{equ:ioeffe-1}
156: \end{equation}
157: where $\kappa\approx40.3$\,m$^3$s$^{-2}$,
158: $c$ is the speed of light (m$\cdot$s$^{-1}$),
159: and $\nu$ the frequency (Hz), and where we neglect magnetic field effects
160: and assume $\nu$ is large compared with the local plasma frequency
161: (for an extreme case, the plasma frequency is of $\sim$15\,MHz).
162: The negative sign applies for phase delays and the positive sign for
163: group delays.
164: In standard astrometric VLBI experiments observations are made simultaneously
165: in two well separated bands of frequencies in order to
166: estimate the ionospheric effect. A nearly vacuum equivalent delay
167: can be obtained from the following expression:
168: \begin{equation}
169: \tau_{\rm free}=\frac{(\nu_1/\nu_2)^2\tau_1 - \tau_2}{(\nu_1/\nu_2)^2-1},
170: \label{equ:iofree-delay}
171: \end{equation}
172: where $\tau_i$ is the delay --either group or phase--
173: at frequency $\nu_i$ ($i=1,2$, $\nu_1>\nu_2$).
174:
175: Thus, with dual frequency observations, the ionosphere effect
176: can largely be removed from the VLBI data. Such
177: removal can also be made for single-frequency observations
178: via Eq.\ (\ref{equ:ioeffe-1}), if
179: estimates of the TEC along the lines of sight of the radio telescopes
180: are available from other observations.
181:
182: Guirado \et\ (\cite{gui95}) showed that it is possible to estimate
183: the ionospheric effect with accuracy useful for astrometric purposes from
184: Faraday-rotation measurements.
185: In this work, the authors used a ``clipped" sinusoidal function to
186: model the diurnal behavior of the TEC.
187: In this model, the night component is
188: constant and equal to the minimum TEC value, and the day component is
189: expressed as the positive part of a sinusoid, with its
190: maximum some hours after noon.
191: Their observations were obtained in late 1985, a
192: time of minimum solar activity.
193:
194: In the method presented here we used
195: GPS measurements that provide
196: TEC values as a function of time.
197: Such GPS-based TEC determinations were first successfully applied to geodetic
198: VLBI by Sard\'on \et\ (\cite{sar92}).
199:
200: \section{The Global Positioning System and the TEC.\label{sec:gps-system}}
201:
202: An introduction to the Global Positioning System (GPS)
203: can be found in Hofmann-Wellenhof \et\ (\cite{hof97}).
204: A main use of the GPS system is
205: to determine the position
206: $(x,y,z,t)$ of a GPS receiver on Earth's surface.
207: The system consists of a constellation of 24 satellites broadcasting
208: electromagnetic signals in two narrow frequency bands,
209: a set of monitoring ground sites, the Master Control Station,
210: and GPS receivers. The 24 satellites orbit the Earth in near
211: circular orbits with a 12\,hr period, at a height of about 20,200\,km,
212: and an inclination of 55\deg. The spacecraft are in six
213: orbital planes with four satellites nearly equally spaced along the orbit
214: in each plane.
215: At any moment, from any point on Earth, it is possible to detect
216: signals simultaneously from 7 to 9 of these satellites.
217: Each satellite broadcasts a block of data
218: every 30 seconds, consisting of a description of its orbit and of
219: GPS time, as well as a pseudo-random code every
220: millisecond (coarse-acquisition C/A, for civilian
221: use) and a precision code (P, 266\,d period, for military use),
222: usable to determine more accurately the position of the ground receiver.
223:
224: The oscillators of the satellites generate a fundamental frequency
225: $\nu_0=10.23$\,MHz, which is the P-code frequency. The C/A-code frequency
226: is $\nu_0/10$. The GPS signal is emitted at two frequencies,
227: 154\,$\nu_0$ and 120\,$\nu_0$ (L1 and L2, respectively, 1,575.42\,MHz and
228: 1,227.60\,MHz, or $\lambda\lambda$19 and 24.4\,cm).
229: L2 carries only the P signal,
230: and L1 both P and C/A signals.
231:
232: \subsection{The GPS observables.\label{sec:gps-obs}}
233:
234: The main GPS observable is
235: $dT$, the time of transit of the signal from the satellite to
236: the ground receiver.
237: The value of $dT$ can be determined, in effect,
238: by comparing the time of broadcasting of
239: the codes P or C/A with the time of receipt of these codes.
240: An observable that corresponds to $dT$
241: can be obtained by cross correlation of the signal
242: received from the satellite at each band with a reference signal
243: generated at the receiver and ``tied" to GPS time.
244: In this case, $dT=\phi/2\pi\nu$, where $\phi$ is the total phase change
245: of the signal during propagation from satellite to receiver and $\nu$ is
246: the center frequency of this transmitted signal.
247: This observable is the
248: phase delay of the carrier signal.
249:
250: The GPS observable is affected by the following: the ionosphere,
251: the troposphere, $2\pi$ phase ambiguities (equivalent to multiples of
252: of 634.75\,ps and 814.60\,ps, respectively, for L1 and L2),
253: multipath (e.g., from signals that
254: are reflected or scattered into the receiver antennas from objects
255: nearby to it), different effective location of
256: receivers for L1 and L2, instrumental delays (different for L1 and L2),
257: degraded coding of the signals, and clock errors.
258: The contributions to the observables of the largest of these effects
259: can be sharply reduced by application of
260: suitable techniques (see, e.g., Blewitt \cite{ble90}).
261:
262: \subsection{Obtaining the TEC from GPS data.\label{sec:gps-equ}}
263:
264: The GPS observable
265: can be modeled as a function of distance from satellite to receiver,
266: ionospheric delays, tropospheric delays, clock
267: errors, and instrumental phase- and group-delay biases
268: (Sard\'on \et\ \cite{sar94}).
269: The ionospheric term can be estimated by a combination of
270: the L1 and L2 observables.
271: We can denote the TEC for any observation direction as $I^i_k(t)$. This TEC
272: is defined along the line of sight from
273: the radio telescope $k$
274: to the radio source $i$ and can be expressed approximately
275: as a function $\cal V$ (vertical
276: value of TEC) of time $t$ and the intersection point $P^i_k$ (``ionospheric
277: point") of the line of
278: sight from $k$ to $i$ with the (average)
279: F$_2$ layer of the ionosphere
280: (at an altitude of $h_{{\rm F}_2}=350$\,km), times
281: the obliquity or slant function $S(e^i_k)$, defined as the secant of the zenith
282: angle at the ionospheric point (see below
283: for geometry clarification), which is a function of the
284: elevation angle $e^i_k$ of the observation:
285: \begin{equation}
286: I^i_k(t)=S(e^i_k)\cdot {\cal V}(P^i_k,t).
287: \label{equ:tec-form}
288: \end{equation}
289:
290: The positions of the involved sites and the ionospheric points
291: can be expressed in a
292: geocentric coordinate system
293: $(X,Y,Z)$ or in a geocentric-solar one $(\Psi,\chi,Z^\prime)$ with the
294: $Z^\prime$ axis directed toward the Sun from the Earth center,
295: $\Psi$ the angle
296: in the $XY$-plane (measured counterclockwise from $X$) and $\chi$ the
297: angle with apex at the Earth's center, measured
298: from the direction to the Sun ($Z^\prime$) to the direction to
299: $P^i_j$.
300: This latter coordinate system is useful
301: since the ionosphere is roughly time-independent
302: in this reference frame.
303: Following Sard\'on \et\ (\cite{sar94}), we replace $\cal V$
304: for a GPS site $j$ and satellites $l$ by its locally
305: linear approximation in $P^l_j$ (ionospheric
306: point towards satellite $l$) using the $(\Psi,\chi,Z^\prime)$-coordinate
307: system:
308: \begin{equation}
309: {I}^l_j(t)=S(e^l_j)\cdot [A_j(t)+B_j(t)d\Psi^l_j(t) +
310: C_j(t)d\chi^l_j] + K^l + K_j.
311: \label{equ:tec-terms}
312: \end{equation}
313: Here, the coefficients for each site are
314: $A$ (0$^{\rm th}$ order, vertical), $B$
315: (1$^{\rm st}$
316: order, $\Psi$-direction), $C$ (1$^{\rm st}$ order,
317: $\chi$-direction). We have also introduced the
318: instrumental
319: GPS satellite $K^l$ and receiver $K_j$ biases.
320: $A_j(t)$ is the
321: vertical TEC at site $j$, $d\Psi^l_j(t)=\Psi^l_j-\Psi_j$, and
322: $d\chi^l_j=\chi^l_j-\chi_j$ are the coordinates
323: $\Psi$ and $\chi$ of the ionospheric point
324: of the line of sight from $j$ towards the satellite $l$ minus
325: the corresponding coordinates
326: of the GPS site $j$.
327: The coefficients $A$, $B$, $C$, can be determined, and the $K$-biases
328: largely removed, using a Kalman filtering method (Herring \et\ \cite{her90})
329: with the data from different satellites $l$ (about 8 at a time) to obtain
330: an estimate of the TEC from GPS data. The $K$ biases
331: can be due to such effects as errors in the estimation of
332: phase ambiguities and multipath (see Sard\'on \et\ \cite{sar94}).
333: In sum, we produce estimates of the total electron content
334: of the ionosphere using data from a network of GPS stations
335: and a multiplicity of satellites.
336:
337: \section{Vertical TEC from GPS and TEC for VLBI.\label{sec:gps2vlbi}}
338:
339: Here we describe the formulas we used to estimate the TEC
340: along the paths of a VLBI observation from the values of the vertical
341: TEC (evaluated by the method described in Sect.\ \ref{sec:gps-equ})
342: at a GPS site near each one of the VLBI sites.
343:
344: Consider a GPS site $j$ at latitude $\zeta_j$ and
345: longitude $\lambda_j$, and a VLBI site $k$ at position
346: $(\zeta_{k},\lambda_{k})$ (see Fig.\ \ref{fig:geom-iono}).
347: The coordinates of $P^i_k$, denoted by $(\zeta_{P^i_{k}},
348: \lambda_{P^i_{k}})$
349: at the epoch the radio source is at elevation $e^i_k$ and azimuth $a^i_k$,
350: are
351: (Klobuchar \cite{klo87}):
352: \begin{eqnarray}
353: \zeta_{P^i_{k}} & = &
354: \arcsin ( \sin \zeta_{k} \cos {\cal E}^i_j +
355: \cos \zeta_{k} \sin {\cal E}^i_j \cos a^i_j)\\
356: \lambda_{P^i_{k}} & = &
357: \lambda_{k} +
358: \arcsin \left(\frac{\sin {\cal E}^i_j \sin a^i_j}{\cos \zeta_{P^i_{k}}}\right),
359: \label{eq:lambdak}
360: \end{eqnarray}
361: where $(-\pi/2 \le \zeta_k \le \pi/2)$ and
362: the $\arcsin$-function in $\lambda$ holds for values in
363: the interval $(-\pi/2,\pi/2)$, appropriate for Eq.\ (\ref{eq:lambdak})
364: since GPS and VLBI sites are nearly collocated.
365: ${\cal E}^i_k$ is the angle, measured from the center
366: of the Earth between the line to site $k$ and the line to the
367: ionospheric point for radio source $i$:
368: ${\cal E}^i_k=\pi/2-e^i_k-\arcsin(\Xi \cos e^i_k)$ (see
369: Fig.\ \ref{fig:slant}, where
370: $\Xi=R_\oplus/(R_\oplus+h_{{\rm F}_2})$, $R_\oplus$ is the Earth's
371: radius, and $\Xi\approx 0.948$ for $h_{{\rm F}_2}=350$\,km).
372: The local time at the
373: ionospheric point is $t=(\lambda_{P^i_{k}}/15)+$UT\,hr ($\lambda$ in
374: degrees, with $\lambda$ positive to the East).
375:
376: \begin{figure*}[htb]
377: \vspace*{308pt}
378: \special{psfile=er_fig1.eps hoffset=30 voffset=0 hscale=80
379: vscale=80}
380: \caption{Geometry used to estimate TEC from GPS measurements (see text).
381: The vertical lines here are radial (from the Earth's
382: center) and the tangent planes drawn
383: are approximations to the surface of spheres (Earth and F$_2$ layer of
384: ionosphere) having, e.g., the VLBI site and its ionospheric
385: point as points of tangency.
386: \label{fig:geom-iono}
387: }
388: \end{figure*}
389:
390: %
391: %
392: \begin{figure}[htbp]
393: %284x169
394: \vspace{135pt}
395: \special{psfile=er_fig2.eps hoffset=0 voffset=0 hscale=80 vscale=80}
396: \caption{Geometry of the obliquity or slant factor, which is the secant of
397: the zenith angle, defined by the elevation $e^i_k$ (see text for definition
398: of $\Xi$). Curvatures
399: of the Earth's surface and the ionosphere
400: F$_2$ layer are exaggerated for clarity.
401: \label{fig:slant}
402: }
403: \end{figure}
404:
405: We assume
406: that the vertical TEC $A_{P^i_{k}}(t)$ at the ionospheric point is
407: given in terms of $A_j(t)$ by:
408: \begin{equation}
409: A_{P^i_{k}}(t)=A_j\left(t+\frac{\lambda_{P^i_{k}} - \lambda_j}{15}\right).
410: \end{equation}
411: This relation effects a longitude correction but ignores any latitude dependence
412: of the TEC values.
413: This approach is reasonable for mid-latitude stations
414: and sources at high declination since GPS sites can be
415: collocated with VLBI sites, or at least placed relatively near to them,
416: and since TEC changes more rapidly with longitude than with latitude.
417:
418: The slant factor was defined above as the secant of the zenith
419: angle, which is $\arcsin (\Xi \cos e^i_k)$ (see Fig.\
420: \ref{fig:slant}, and Klobuchar \cite{klo87}).
421: Thus, the TEC at the ionospheric point $P^i_{k}$, mapped by
422: the slant factor (Eq.\ \ref{equ:tec-form}) gives
423: \begin{equation}
424: I^i_{k} (t)= \sec\{\arcsin (\Xi \cos e^i_k)\} \cdot
425: A_{P^i_{k}}(t),
426: \end{equation}
427: which, with Eq.\ (\ref{equ:ioeffe-1}) yields our estimate
428: of the ionospheric delays at the indicated VLBI site. The overall ionospheric
429: effect on a VLBI observable is a simple linear combination of the effects from
430: each of the two sites involved in the observation.
431:
432:
433: \section{A case to test the method: VLBI observations
434: of the Draco triangle.}
435:
436: Progress in high precision phase-delay difference
437: astrometry has been made by Ros \et\ (\cite{ros99}) through VLBI
438: observations of triangle formed by the radio sources
439: \object{BL\,1803+784}, \object{QSO\,1928+738} and \object{BL\,2007+777},
440: in the Northern
441: constellation of Draco (the Dragon). The observations were made
442: simultaneously at the frequencies of 2.3 and 8.4\,GHz
443: at epoch 1991.89 with an intercontinental interferometric array.
444: The angular separations among these radio sources were determined
445: with submilliarcsecond accuracy from a weighted-least-squares
446: analysis of the differenced and undifferenced phase delays.
447: The modeling of these
448: astrometric VLBI observations was sufficiently accurate to
449: estimate reliably the ``$2\pi$ ambiguities" in the differenced phase delays for
450: source separations of
451: almost 7$^\circ$ on the sky. For such angular distances,
452: this accurate ``phase connection"
453: at 8.4\,GHz, yielding the phase delays with standard errors
454: well within one
455: phase cycle over the entire session of observations,
456: was demonstrated
457: at an epoch of solar maximum.
458: As in earlier works (e.g., Guirado \et\ \cite{gui95}, Lara \et\
459: \cite{lar96}), after phase connection
460: the effects of the extended
461: structure of the radio sources were largely removed from the phase-delay
462: observables. The effects
463: of the ionosphere
464: were also mostly removed, via the GPS-based method described in this paper.
465:
466: In 1991 the number of available GPS sites was small, and only
467: data from Goldstone and Pinyon Flats in the US, and from
468: Herstmonceux and Wettzell in Europe (see Table \ref{tab:gps-vlbi} for
469: details) were available to be used for our experiment.
470: GPS data from the two US sites were
471: used for the VLBI sites at
472: Fort Davis (TX), Pie Town (NM), Kitt Peak
473: (AZ), and Los Alamos (NM) to estimate the TEC for observations at these
474: sites.
475: Similarly, GPS data from both European sites were used for
476: Effelsberg (Germany).
477: The VLBI observations were carried out from 14\,hr\,UT
478: on 20 November to 4\,hr\,UT on 21 November 1991. In Fig.\
479: \ref{fig:gm6-tec} we
480: show the corresponding vertical TEC values for one of the
481: GPS sites on each continent.
482: We see the dusk and night part of the data for Wettzell, and
483: the daylight data for Pinyon Flats.
484:
485: \begin{table}[htb]
486: \caption[]{Positions of the GPS sites and the VLBI sites (Ros \et\ \cite{ros99})
487: and maximum and minimum values of the vertical
488: TEC value on 20/21 November 1991 (see text for definitions of
489: notation).\label{tab:gps-vlbi}}
490: \begin{tabular}{@{}l@{~}r@{~}r@{~}c@{~~}c@{~}l@{~}r@{~}r@{}}\hline
491: GPS & $\zeta_j$ & $\lambda_j$
492: & $A_j^{\rm min}$
493: & $A_j^{\rm max}$
494: & VLBI & $\zeta_{k}$
495: & $\lambda_{k}$ \\
496: site & & & {\tiny [TECU]}
497: & {\tiny [TECU]}
498: & site & & \\ \hline
499: {\tt DS10}$^1$ & 35.2\,N & 116.9\,W & 5 & 57 & {\tt FD}$^a$& 30.6\,N & 103.9\,W \\
500: {\tt PIN1}$^2$ & 33.6\,N & 116.5\,W & 5 & 57 & {\tt PT}$^b$& 34.3\,N & 108.1\,W \\
501: & & & & & {\tt KP}$^c$& 32.0\,N & 111.6\,W \\
502: & & & & & {\tt LA}$^d$& 35.8\,N & 106.2\,W \\ \hline
503: {\tt HERS}$^3$ & 50.9\,N & 0.3\,E & 8 & 38 & {\tt EB}$^e$& 50.3\,N & 6.8\,E \\
504: {\tt WTZ1}$^4$ & 49.1\,N & 12.8\,E & 8 & 39 & & & \\ \hline
505: \end{tabular}
506: \noindent
507: \\
508: {\footnotesize
509: $^1$~Goldstone Deep Space Tracking Station, NASA, CA, US;
510: $^2$~Pinyon Flats Observatory, SIO, UCSD, CA, US;
511: $^3$~Herstmonceux, East Sussex, England;
512: $^4$~K\"otzting, IFAG, BKG, Bavaria, Germany.\\
513: $^a$~Fort Davis, VLBA, NRAO, TX, US;
514: $^b$~Pie Town, VLBA, NRAO, NM, US;
515: $^c$~Kitt Peak, VLBA, NRAO, AZ, US;
516: $^d$~Los Alamos, VLBA, NRAO, NM, US;
517: $^e$~Effelsberg, MPIfR, North-Rhine-Westphalia, Germany.
518: }
519: \end{table}
520:
521: \begin{figure}[htb]
522: \vspace*{225pt}
523: \special{psfile=er_fig3.ps hoffset=-10 voffset=-20 hscale=47
524: vscale=47}
525: \caption{Values of the vertical
526: TEC at the GPS sites of Pinyon Flats (CA, US) and
527: Wettzell (Germany)
528: (Ros et al.\ 1999). These values were used to compute the TECs
529: at nearby VLBI sites.
530: We assumed an error
531: of 1.5\,TECU (represented by the dotted lines), much
532: larger than the statistical standard errors.
533: \label{fig:gm6-tec}}
534: \end{figure}
535:
536: From the GPS-based estimates of the TEC along the lines of sight
537: for each VLBI site, we
538: calculated the ionospheric contribution for each baseline and epoch
539: of observation.
540: These ionospheric contributions were removed from the VLBI observables.
541: For the phase delays, these contributions
542: ranged in magnitude from 0 to 1.2\,ns
543: at 8.4\,GHz for intercontinental baselines, and were less than 0.1\,ns for
544: continental baselines. The intercontinental baseline lengths range
545: from $\sim$7800 to
546: $\sim$8300\,km, and the differences in local
547: time are $\sim$7.5\,hr
548: (see Fig.\ \ref{fig:gm6-tec}). Since the ionospheric
549: effect is the combination of effects for both antennas, the ionospheric
550: delay is quite important
551: for our intercontinental baselines.
552: By contrast our US-continental baseline lengths
553: range from $\sim$200 to $\sim$750\,km which corresponds at most to
554: about 30 minutes
555: difference in local time and a much smaller ionospheric effect on
556: the VLBI observables.
557:
558: Eq.\ (\ref{equ:iofree-delay}) provides the usual way to largely
559: remove the effect of the ionosphere
560: on dual-band VLBI observations.
561: The phase delays from our 2.3\,GHz observations
562: could not be freed from $2\pi$ ambiguities due to the large
563: scatter in these data.
564: However, unambiguous but
565: less precise group delays were available at both 2.3 and
566: 8.4\,GHz. The ionospheric contributions were estimated from
567: these group-delay data. In Fig.\ \ref{fig:gm6-gps-gde} we
568: show the comparison
569: of GPS-based and dual-band VLBI-based ionospheric delay estimates
570: at 8.4\,GHz.
571: We show four of the ten available baselines as representative examples:
572: two intercontinental and two continental-US ones.
573: Apart from the larger dispersion in the
574: group-delay than in the GPS-based data,
575: this comparison provides a good independent confirmation of
576: the reliability of the GPS-based method for the correction of VLBI data.
577: The error bars for the group delays shown in the
578: figure are the appropriate combination of the
579: statistical standard errors for the data at these frequencies
580: (see Eq.\ \ref{equ:iofree-delay}).
581: The statistical standard errors for the GPS
582: estimates are each about 0.2\,TECU. We
583: assume a much larger standard error --1.5\,TECU-- to try to account
584: for possible inaccuracies not estimated with the Kalman filtering, such
585: as incorrect values for $h_{{\rm F}_2}$, the consequent changes in
586: the mapping function and in the eventual position of the ionospheric
587: point, and other unmodeled effects. Thus, we infer
588: a corresponding contribution of $\sim$30\,ps to the standard errors of the
589: phase delays at 8.4\,GHz.
590: For the data presented in Fig.\ \ref{fig:gm6-gps-gde},
591: the root-mean-square
592: of the differences between
593: the results from the two
594: methods are below 0.15\,ns for intercontinental baselines and
595: 0.10\,ns for continental ones.
596:
597: \begin{figure}[htb]
598: \vspace*{260pt}
599: \special{psfile=er_fig4.ps hoffset=-25 voffset=-30 hscale=50
600: vscale=50}
601: \caption{Comparison between the ionospheric delays at 8.4\,GHz
602: estimated by the combination of group delays at 8.4 and 2.3\,GHz from
603: the VLBI observations of Ros et al. (\cite{ros99}) (points with error bars)
604: and the corresponding ionospheric delays
605: estimated by the GPS TEC determination reported here (solid line).
606: Data from four representative baselines of a total of ten from the
607: VLBI experiment are shown here. Root-mean-square values for the
608: differences are, respectively, of 0.11, 0.15, for the upper panels
609: and of 0.10, 0.08\,ns for the lower panels.
610: \label{fig:gm6-gps-gde}}
611: \end{figure}
612:
613: \section{Conclusions}
614:
615: Assuming that the ionosphere can be modeled usefully as a thin
616: shell surrounding the Earth at a height of 350\,km,
617: we estimate the TEC for the line of sight from the GPS site to the
618: satellite using ground reception of GPS signals and knowledge of the orbital
619: parameters of the GPS satellites.
620: Having a GPS
621: site near a VLBI site, we can reliably ``transfer" this estimate to the TEC
622: for the line of sight from the radio telescope to the radio source.
623: These TEC values allow us to correct the
624: VLBI observables for the effects of the ionosphere for any frequency.
625: We made such corrections for our phase-delay
626: data at 8.4\,GHz (Ros \et\ \cite{ros99}) for
627: an epoch at which there was a paucity of relevant GPS data.
628: The estimates of the ionospheric delays provided
629: by the VLBI measurements of group delay
630: at 2.3 and 8.4\,GHz differ from the corresponding
631: delays obtained from GPS data to within
632: root-mean-square values
633: below 0.15\,ns for intercontinental baselines and
634: 0.10\,ns for continental ones.
635: Thus, we have shown, in particular, that the GPS determination of TEC
636: can be successfully used in the astrometric analysis of VLBI observations.
637:
638: The density of the network of GPS sites has increased dramatically
639: since 1991 and the accuracy of the TEC deduced from GPS data has
640: improved significantly. Given such progress, the approximations
641: used in this paper are no longer necessary.
642: Now GPS estimates for the vertical TEC are available from virtually every
643: land location all of the time and thus for every VLBI observation.
644:
645: The advantages of GPS compared with geostationary
646: beacons which used Faraday rotation to determine TECs are notable:
647: global land coverage for GPS is available from
648: geodetic networks (e.g., the International
649: GPS Geodynamics Service);
650: the TEC estimates do not depend on assumptions about the Earth's
651: magnetic field; L1 and L2 GPS data are available over
652: the internet in standard formats (e.g., RINEX:
653: Receiver INdependent EXchange).
654: We also note that the
655: ionosphere cannot always be represented by a thin shell model with
656: good accuracy; %, but such a model is not so important when GPS receivers
657: %are collocated with VLBI sites;
658: moreover, more accurate models
659: can be devised and suitably
660: parameterized given the tomography-like sampling of the ionosphere
661: provided by the GPS. The biases in the GPS observables have
662: to be properly corrected in the estimates of TECs.
663: Davies \& Hartmann (\cite{dav97}) set an upper limit of 3\,TECU for the
664: present agreement of GPS TEC with the results from other methods,
665: but only of few 0.01\,TECU for the relative errors of TEC estimates
666: over the course of some hours for any given site.
667: The latter represents relative changes
668: of delay at 8.4\,GHz of 0.2\,ps, nearly two orders of magnitude smaller than
669: the delay equivalent of
670: a $2\pi$ phase ambiguity. With such accuracies and the present network
671: of GPS sites, the removal of most of the effect of the ionosphere
672: from VLBI observations should not be difficult, although much of the
673: smaller-scale ionospheric activity cannot be adequately sampled by GPS.
674: From dual-band difference VLBI astrometry
675: and GPS data, the optical depth of the emission in
676: radio sources can be better
677: studied by comparing brightness distributions
678: obtained independently at each frequency band with respect to the
679: same coordinate system.
680: In sum, the introduction of GPS techniques should greatly improve
681: the scientific results obtained from VLBI observations.
682:
683: \acknowledgements E.R.\ acknowledges a F.P.I.\ fellowship of the
684: Generalitat Valenciana.
685: We acknowledge the referee, Dr.\ R.M.\ Campbell for his very helpful
686: suggestions and remarks.
687: We are grateful to the SOPAC/IGPP, at SIO, University of California,
688: San Diego (US), for kindly providing the GPS data from their GARNER archives,
689: and to Prof.\ R.T.\ Schilizzi for encouragement.
690: This work has been partially supported by the Spanish DGICYT grants
691: PB\,89-0009, PB\,93-0030, and PB\,96-0782, and by the U.S.\ National
692: Science Foundation Grant No.\ AST 89-02087.
693:
694: \begin{thebibliography}{}
695:
696: \bibitem[1990]{ble90}
697: Blewitt G., 1990,
698: %`An automatic editing algorithm for GPS data', {\it
699: Geophys.\ Res.\ Lett., 17 (3), 199%--202
700:
701: \bibitem[1997]{dav97}
702: Davies K., Hartmann G.K., 1997,
703: %`Studying the ionosphere with the Global Positioning System',
704: Radio Science 32, 1695%--1703
705:
706: \bibitem[1995]{gui95}
707: Guirado J.C., Marcaide J.M., El\'osegui P., \et\,
708: %Ratner, M.I.,
709: %Shapiro, I.I., Eckart, A., Quirrenbach, A., Schalinski, C.J., Witzel, A.,
710: 1995,
711: %`VLBI Differential Astrometry of the Radio Source 1928+738 and
712: %2007+777 at 5 GHz',
713: \aap\ 293, 613%--625
714:
715: \bibitem[1990]{her90}
716: Herring T.A., Davis J.L., Shapiro I.I., 1990,
717: %`Geodesy by Radio Interferometry: The Application of Kalman
718: %Filtering to the Analysis of Very Long Baseline Interferometry Data',
719: \jgr\ B, 95, 12561%--12581
720:
721: \bibitem[1997]{hof97}
722: Hofmann-Wellenhof B., Lichtenegger H., Collins J., 1997,
723: Global Positioning System: Theory and Practice, Springer Verlag,
724: Vienna, Austria
725:
726: %\bibitem[1975]{klo75}
727: %Klobuchar J.A., 1975,
728: %%`A First-Order, Worldwide, Ionospheric, Time-Delay Algorithm',
729: %Air Force Cambridge Research Laboratories Report No.~AFCRL-TR-75-0502
730: %(NTIS ADA 018862)
731:
732: \bibitem[1987]{klo87}
733: Klobuchar, J.A., 1987,
734: IEEE Trans.\ on Aerospace \& Electronic Systems AES-23, 325
735: %``Ionospheric Time-Delay Algorithm for Single-Frequency GPS Users"
736:
737: \bibitem[1996]{lar96} Lara, L.,
738: Marcaide J.M., Alberdi A., Guirado J.C., 1996,
739: %{\em VLBI Differential Astrometry at Large Angular Separation
740: %3C\,395-3C\,382},
741: \aap\ 314, 672%--678
742:
743:
744: %\bibitem[\protect\citeauthoryear{Kovalevsky}{1995}]{kov95}
745: %Kovalevsky, J.: 1995, {\it Modern Astrometry}, Springer Verlag
746:
747: %\bibitem[\protect\citeauthoryear{Leick}{1990}]{lei90}
748: %Leick, A.: 1990, {\it GPS satellite surveying}, Wiley Interscience
749:
750: \bibitem[1999]{ros99}
751: Ros E., Marcaide J.M., Guirado J.C., \et\,
752: %Ratner, M.I., Shapiro, I.I.,
753: %Krichbaum, T.P., Witzel, A., Preston, R.A.:
754: 1999,
755: %`High Precision Differential Astrometry on the triplet of S5 Radio
756: %Sources 1803+784/ 1928+738/ 2007+777',
757: \aap\ 348, 381
758:
759: \bibitem[1992]{sar92}
760: Sard\'on E., Rius A., Zarraoa N., 1992,
761: %`Ionospheric calibration of VLBI data using GPS observations',
762: Proc.\ %eedings of the
763: Symp.\ %osium on
764: Refraction of Transatmospheric Signals in Geodesy, Netherlands
765: Geodetic Comision, Publications on Geodesy No.\ 36, p.\ 59%--62
766:
767: \bibitem[1994]{sar94}
768: Sard\'on E., Rius A., Zarraoa N., 1994,
769: %`Estimation of the transmitter and receiver differential biases and
770: %the ionospheric total electron content from Global Positioning System
771: %observations',
772: \rsci, 29, 577%--586
773:
774: %\bibitem[1953]{smi53}
775: %Smith, E.K., Weintraub, S., 1953,
776: %Proceedings of the IRE, 41, 1035%--1037
777: %%Smith, E.K., Weintraub, S.: 1953, `The Constants in the
778: %%Equation for Atmospheric Refractive Index at Radio Frequencies',
779: %%{\it Proceedings of the IRE}, {\bf Vol.\ no.\ 41}, pp.\ 1035--1037
780:
781: %\bibitem[1994]{sov94}
782: %Sovers, O.J., Jacobs, C.S., 1994,
783: %%`Observation Model and Parameter
784: %%Partials for the JPL VLBI Parameter Estimation Software
785: %%``MODEST"--1994', {\it
786: %JPL Publication, 83-39, Rev.~5
787:
788: \bibitem[1986]{tho86}
789: Thompson A.R., Moran J.M., Swenson G.W., 1986,
790: Interferometry and Synthesis in Radio Astronomy, J.\ Wiley and Sons,
791: New York, NY, US
792:
793: \end{thebibliography}
794:
795: \end{document}
796:
797: \bibitem[\protect\citeauthoryear{Alef}{1989}]{ale89}
798: Alef, W., `Scheduling, Correlating, and
799: Postprocessing of VLBI Observations', in {\it Very Long
800: Baseline Interferometry. Techniques and Applications}, eds.\
801: M.~Felli y R.E.~Spencer, Kluwer, Dordrecht, The Netherlands, 97--139 (1989)
802:
803: %\bibitem[\protect\citeauthoryear{Blewitt}{1990}]{ble90}
804: %Blewitt, G., `An automatic editing algorithm for GPS data',
805: %{\it Geophysical}\\
806: %{\it Research Letters}, 17 (3), 199--202
807: %(1990)
808:
809: \bibitem[\protect\citeauthoryear{Chao}{1972}]{cha72}
810: Chao, C.C., `A model for tropospheric calibration from daily surface
811: and radiosonde balloon measurements', {\it Technical Memo
812: California Institute of Technology-}\\ {\it Jet
813: Propulsion Laboratory}, {\bf 391--350}, 17pp. (1972)
814:
815: \bibitem[\protect\citeauthoryear{Chao}{1974}]{cha74}
816: Chao, C.C., `The troposphere calibration model for Mariner
817: Mars 1971', {\it JPL/NASA Technical Report}, {\bf 32-1587}, 61
818: (1974)
819:
820: \bibitem[\protect\citeauthoryear{Davis et al.}{1985}]{dav85}
821: Davis, J.L., Herring, T.A., Shapiro, I.I., Rogers, A.E.E., Elgered, G.,
822: `Geodesy by radio interferometry - Effects of atmospheric modeling
823: errors on estimates of baseline length', \rsci, {\bf 20}, 1593--1607
824: (1985)
825:
826:
827: \bibitem[\protect\citeauthoryear{Guirado et al.}{1995b}]{gui95b}
828: Guirado, J.C.,
829: Marcaide, J.M., Alberdi, A., El\'osegui, P.,
830: Ratner, M.I., Shapiro, I.I., Kilger, R., Mantovani, F., Venturi, T.,
831: Rius, A., Ros, E., Trigilio, C., Whitney, A.R., `Proper Motion of
832: Components in 4C39.25', \aj, {\bf 110}, 2586--2596 (1995)
833:
834: \bibitem[\protect\citeauthoryear{Guirado et al.}{1998}]{gui98}
835: Guirado, J.C., Marcaide, J.M., Ros, E., Ratner, M.I.,
836: Shapiro, I.I., Quirrenbach, A., Witzel, A.,
837: {\em Submilliarcsecond shift of the brightness peak of
838: the radio sources 1928+738 and 2007+777}
839: {\it Astronomy and Astrophysics, en preparaci'on (1998)
840:
841: \bibitem[\protect\citeauthoryear{Herring}{1983}]{her83}
842: Herring, T.A., {\it Precision and Accuracy of
843: Intercontinental distance Determinations
844: using Radio Interferometry}, PhD Thesis, MIT, Cambridge
845: MA, EE.UU. (1983)
846:
847: \bibitem[\protect\citeauthoryear{Herring et al.}{1990}]{her90}
848: Herring, T.A., Davis, J.L., Shapiro, I.I.,
849: {\em Geodesy by Radio Interferometry: The Application of Kalman
850: Filtering to the Analysis of Very Long Baseline Interferometry Data},
851: \jgr, {\bf 95}, 12561--12581 (1990)
852:
853: \bibitem[\protect\citeauthoryear{IERS}{1993}]{IERS94}
854: {\it 1993 IERS Annual Report}, International Earth Rotation
855: Service, Paris (1994)
856:
857:
858: %\bibitem[\protect\citeauthoryear{Kovalevsky}{1995}]{kov95} Kovalevsky, J., {\it Modern
859: %Astrometry}, Springer Verlag, Heidelberg (1995)
860:
861: \bibitem[\protect\citeauthoryear{Kraus}{1986}]{kra86} Kraus, J.D.,
862: {\it Radioastronomy, 2nd edition}, Cygnus-Quasar
863: Books, Powell (Ohio) (1986)
864:
865: %\bibitem[\protect\citeauthoryear{Leick}{1990}]{lei90}
866: %Leick, A., {\it GPS satellite surveying}, Wiley Interscience
867: %(1990)
868:
869: \bibitem[\protect\citeauthoryear{Robertson}{1975}]{rob75}
870: Robertson, D.S., {\it Geodetic
871: and Astrometric Measurements with Very-Long-}\\
872: {\it Baseline Interferometry}, PhD Thesis,
873: Massachusetts Institute
874: of Technology, Cambridge, MA, EE.UU. (1975)
875:
876: \bibitem[\protect\citeauthoryear{Rogers et al.}{1983}]{rog83} Rogers, A.E.E., Capallo, R.J., Hinteregger,
877: H.F., Levine, J.I., Nesman, E.F., Webber, J.C., Whitney, A.R., Clark,
878: T.A., Ma, C., Ryan, J., Corey, B.E., Counselman III, C.C., Herring, T.A.,
879: Shapiro, I.I., Knight, C.A., Shaffer, D.B., Vandenberg, N.R., Lacasse,
880: R., Mauzy, R., Rayhrer, B., Schupler, B.R., Pigg, J.C.,
881: {\em Very-Long-Baseline Interferometry: The Mark III System for Geodesy,
882: Astrometry and Aperture Synthesis}, \sci, {\bf 219},
883: 51--54 (1983)
884:
885: \bibitem[\protect\citeauthoryear{Sard'on}{1993}]{sar93}
886: Sard'on,~E., {\it Calibraciones ionosf'ericas en geodesia
887: espacial mediante el uso de}\\
888: {\it datos GPS}, Tesis Doctoral,
889: Instituto de Astronom'ia y Geodesia, Madrid (1993)
890:
891: \bibitem[\protect\citeauthoryear{Sard'on et al.}{1994}]{sar94}
892: Sard\'on, E., Rius, A., Zarraoa, N.,
893: {\em Estimation of the transmitter and receiver differential biases and
894: the ionospheric total electron content from Global Positioning System
895: observations}, \rsci, {\bf 29}, 577--586 (1994)
896:
897: \bibitem[\protect\citeauthoryear{Shapiro et al.}{1979}]{sha79}
898: Shapiro, I.I., Wittels, J.J., Counselman
899: III, C.C., Robertson, D.S., Whitney, A.R., Hinteregger, H.F., Knight,
900: C.A, Rogers, A.E.E., Clark, T.A., Hutton, L.K., Niell, A.E., {\em
901: Submilliarcsecond Astrometry via VLBI. I. Relative Position of the Radio
902: Sources 3C\,345 and NRAO\,512}, \aj, {\bf 84}, 1459--1469 (1979)
903:
904:
905: