Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.gao.spb.ru/english/personal/malkin/publ/zm13o.pdf Äàòà èçìåíåíèÿ: Fri Nov 8 18:01:00 2013 Äàòà èíäåêñèðîâàíèÿ: Thu Feb 27 23:21:44 2014 Êîäèðîâêà: ISO8859-5 Ïîèñêîâûå ñëîâà: ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï

Searching for an Optimal Strategy to Intensify Observations of the Southern ICRF sources in the framework of the regular IVS observing programs
? ? ÄÄ Zinovy Malkin, Jing Sun, Johannes Bohm, Sigrid Bohm, Hana Krasna

Abstract The quality of the VLBI-derived ICRF in the southern hemisphere is much worse than in the northern hemisphere. The main reason is that only about 3% of the observations have been made of the sources at declinations below -30 deg due to the relatively small number of VLBI stations located in the southern countries. In this paper, we investigated a possibility to increase the number of observations of the existing and prospective southern ICRF radio sources by inclusion of more such sources in the regular IVS sessions like R1 and R4. We tested the influence of adding supplementary southern sources to the IVS R1541 (12JUL09XA) session on EOP and baseline length repeatability with Monte Carlo simulations. We found that adding more observations of southern sources to the standard schedule causes a slight degradation of some geodetic products and a slight improvement of others, depending on the number of added southern sources. Similar results were obtained for the IVS R1591 (13JUN24XA) session. Generally, it has been shown that it is possible to increase the number of observations of southern sources without loss of the overall accuracy of geodetic products. So, the task is to find an optimal trade-off between the maximum increasing of the number of observations of southern sources and the degradation of geodetic results.

1 Introduction
The quality of the ICRF in the southern hemisphere is much worse than in the northern hemisphere. The main reason is that the number of southern VLBI stations participating in the astrometric observing programs is much smaller than that in the northern hemisphere. As a consequence, the number of observations of the southern sources is very small. Only about 3% of the observations have been made of the sources at declinations below -30 deg (see Fig. 1). The situation improves with time, but very slowly despite new southern stations and new CRF-dedicated observing programs (see Fig. 2). The relative number of observations of most southern sources does not improve with time at all. Deficiency of observations of southern sources leads to the following well recognized consequences: Ç the number of the southern ICRF sources is much smaller than the northern; Ç the number of the southern ICRF sources with reliable position and stability estimate, herein reliable core/defining sources, is much smaller than the northern; Ç the position accuracy of the southern sources is generally worse than the northern. Special CRF programs for the southern hemisphere are rare, and are often conducted on poor networks of 2-3 stations, which can deteriorate the source position accuracy because of the source structure effect. Two possible ways were proposed by (Malkin et al., 2012) to increase the number of observations of poorly observed and new prospective ICRF sources on the southern sky: inclusion of more such sources in the regular IVS sessions like R1 and R4, and implementing new scheduling strategies not requiring sky coverage for the
1

Keywords VLBI, IVS, ICRF, scheduling
Zinovy Malkin Pulkovo Observatory and St. Petersburg State University, Pulkovskoe Sh. 65, St. Petersburg 196140, Russia ? ? ÄÄ Jing Sun, Johannes Bohm, Sigrid Bohm, Hana Krasna Department of Geodesy and Geoinformation E120/4, Vienna University of Technology, Guïhausstraïe 27-29, 1040 Vienna, Austria


2

Malkin et al.

stations. In this paper, we investigate possible strategies to force an improvement in the ICRF sources observation distribution over the sky by: Ç including prospective ICRF sources in the regular IVS observing programs, such as R1 and R4; Ç finding a trade-off between a slight degradation of the EOP precision and the long-term ICRF improvement. We made use of the VieVS scheduling and simula? tion tools (Bohm et al., 2012) for our study.
40

DE < -30 DE < -60 3

2

1

0 1990

1995

2000 Year

2005

2010

Fig. 2 Percentage of the observations of southern sources (cu-

mulative by date).

30

ipated. The southern network size ensures large common view, and the multi-baseline observations are important to mitigate the source structure effects.

20

10

0 -90 40

-60

-30

0 Declination, deg

30

60

90

30

20

10

Fig. 3 11 stations network of IVS R1541 session, 5 out of 11

stations are located in the southern hemisphere.
0 -90 -60 -30 0 Declination, deg 30 60 90

Fig. 1 Percentage of observations by DE bands (top) and per-

centage of the well observed sources with Nsess 10, Nobs 200 (bottom). Actual numbers of observations are shown by grey boxes, numbers of observations expected for a uniform distribution are shown by thick lines).

2.1 Scheduling
The original schedule for the R1541 IVS session was generated making use of the SKED software (Gipson, 2010). There are 60 sources observed, 7 southern sources having declination less than -40 degrees. For comparisons, the supplementary southern sources are added to the original source list and experimental schedules are obtained to evaluate the trade-off between the number of southern sources and the accuracy of geodetic products. Considering all the ICRF2 sources having the declination less than -40 degrees, they are sorted by some generalized criteria involving number of sessions, number of observations, and position uncertainty.

2 Monte Carlo Simulation
The IVS R1541 (12JUL09XA) session was used for the Monte Carlo simulations in this paper. The R1541 session network includes 11 stations, 5 of them are located in the southern hemisphere (see Fig. 3). As expected, the Auscope (Australian VLBI Network), station Hartrao, station Tigo, and station Fortleza partic-


A Way to Intensify Obser vations of Southern Sources

3
20

"The worst end" of the list shows which sources we should consider first. The strong southern sources have preference in this study. Schedule 'R1' is achieved with the original source list. Schedule 'R1+' includes three more southern sources and schedule 'R1++' includes six more southern sources as compared with the original R1541 schedule. The three schedules for 24-hour continuous observations are generated with VieVS scheduling package (Sun et al., 2011). The distribution of observed sources is shown in Fig. 4, and detailed information on southern sources is given in Table 1.
Table 1 Number of scans/observations of southern sources in the

15

10

5

0

-90

-60

-30

0 Declination, deg

30

60

90

20

15

IVS R1541 (R1) and two experimental schedules R1+ and R1++. Source 0637-752 0537-441 1104-445 2052-474 2300-683 0048-427 0308-611 2232-488 2204-540 2142-758 0208-512 0332-403 1424-418 Total R1 39 / 39 55 / 88 16 / 18 42 / 48 3/ 3 4/ 6 4/ 4 R1+ 42 / 48 56 / 91 25 / 27 49 / 57 1/ 1 7 / 11 6/ 6 7/ 7 9/ 9 7/ 7 R1++ 37 / 39 59 / 102 19 / 23 46 / 50 1/ 1 7/ 7 2/ 2 3/ 3 6/ 6 3/ 3 18 / 18 47 / 82 42 / 67 290 / 403
10

5

0

-90

-60

-30

0 Declination, deg

30

60

90

20

15

178 / 295

209 / 264

10

5

Except the different source list, the basic scheduling settings used in VieVS are in correspondence with the original R1541 schedule as summarized below. The optimization of source-based strategy is employed with VieVS for this study. Ç Ç Ç Ç Ç Ç frequency setup: R1 frequency setup (X/S band) SNR: 20/15 (15/12 for Tigo) recording data rate: 256 Mbps cut-off elevation angle: 5 degrees minimum scan length: 40 seconds extra time for settling down, calibration, correlator synchronizing

0

-90

-60

-30

0 Declination, deg

30

60

90

Fig. 4 Distribution of observed sources in the original R1541

schedule (top) and two experimental schedules: R1+ (middle) and R1++ (bottom).

2.2 Simulating
For the Monte Carlo simulations, 50 sessions were simulated using the same 24-hour schedule but different

realizations of noise delays, each time creating new values for wet zenith delay, clocks and white noise to simulate observations as realistic as possible. The random errors in delay measurement were modelled by white noise with given power spectral density (PSD). The clock rate instability was modelled using the Allan standard deviation (ASD). The turbulent troposphere was modelled using the site-dependent structure constant Cn , effective wet height H , and wind velocity V . The simulation parameters are summarized in Tables 2


4

Malkin et al. Table 4 Repeatability and standard deviation of EOP for the IVS

and 3). See Sun et al., 2011 for details of the stochastic models used during simulation.
Table 2 Simulation parameters.

R1541 and two experimental schedules R1+ and R1++. Parameter Number of scans Number of observations EOP repeatability [Å as, Å s] R1 1258 3905 143.2 98.2 5.6 36.2 45.0 94.8 77.2 4.4 29.8 29.1 R1+ 1351 3813 125.5 79.1 4.6 42.8 39.5 95.6 77.3 4.6 30.9 29.6 R1++ 1375 3997 98.2 96.8 5.9 39.1 37.2 93.4 74.8 4.7 29.5 28.1

H [m] Vn [m/s] Ve [m/s] wzd0 [mm] d hseg [h] d h [m] clock ASD WN PSD [ps]

2000 0.00 8.00 250 2 200 10-14 @50 min 32

Mean EOP uncertainty [Å as, Å s]

Xp Yp UT1 dX dY Xp Yp UT1 dX dY

Table 3 Site-dependent constant Cn , m-1/3 .

10

Sta name NYALES20 ONSALA60 TSUKUB32 WESTFORD WETTZELL YEBES40M KOKEE

Cn Ç 10 0.65 2.19 3.45 2.30 1.50 1.48 1.39

7

Sta name HARTRAO KATH12M TIGOCONC WARK12M YARRA12M FORTLEZA

Cn Ç 10 1.34 1.68 2.08 1.94 1.76 2.46

7

0 -10 0 10 20 30 40 50

10 0 -10 0 10 20 30 Baseline length index 40 50

Fig. 5 Differences in baseline length repeatability [mm] between

3 Results
The simulated NGS data files are entered into the software package VieVS, which computes a classical least squares solution. All the source coordinates were fixed to the ICRF2 positions (Ma et al., 2009). The standard deviation of the 50 EOP estimates and mean formal uncertainties are listed in Table 4. Fig. 5 shows baseline length repeatability obtained from the simulations. For the baselines shorter than 5,000 km the R1 schedule shows the best result, and R1+ and R1++ schedules yield worse repeatability, whereas for longer baselines the R1++ schedule is the best, and R1 is the worst. However, in fact, the results obtained with the three schedules are close to each other. The mean baseline length repeatability derived from R1, R1+, and R1++ schedules are 13.5 mm, 12.4 mm, and 11.9 mm, respectively. It has been found that further increasing of the number of southern sources (cf. R++ and R+ schedules) leads to a small degradation of baseline length repeatability for short baselines, and small improvement for

two schedules: R1+ minus R1 (top) and R1++ minus R1 (bottom). The horizontal axis represents the 55 baselines with the shortest one WETTZELL-YEBES40M (1575 km) on the left and the longest one TIGOCONC-TSUKUB32 (12401 km) on the right.

long baselines. Errors in some EOP become smaller with inclusion of more southern sources, and some EOP show small degradation in the accuracy.

4 Summary
Including more southern sources in the regular IVS sessions may be a practical way to force an improvement of the VLBI-based ICRF in the southern hemisphere. In this paper, we studied a trade-off between the small degradation of the EOP precision and the ICRF improvement using the source-based scheduling algorithm (Sun et al., 2011). We found no degradation of the overall accuracy of main geodetic products, such


A Way to Intensify Obser vations of Southern Sources

5

as EOP and baseline length repeatability after the inclusion of several supplementary southern sources. Although the number of southern sources added and the number of their observations are not large with respect to the standard scheduling algorithm, regular inclusion of selected sources needed for densification and accuracy improvement of the ICRF in the southern hemisphere will provide a valuable contribution to the next VLBI-based ICRF. Having quarterly observations during two years will give us a good preliminary estimate of both average source position and its stability. Rotating the list of supplementary sources between sessions, e.g., on the quarterly basis, we could substantially increase the number of reliably observed southern sources. The latter is, in particular, very important for selection of new ICRF core (defining) sources. We tested a new approach to the scheduling using two IVS sessions R1541 (12JUL09XA) and R1591 (13JUN24XA). The results obtained with the first session are described in this paper in detail; the results obtained with the second session are very similar. However, a serious problem for schedule optimization is that southern stations are equipped with relatively small antennas, which makes it difficult to observe the weak sources. However, the much greater recording rate (as compared with the present R1/R4 operations) planned for the VLBI2010 observation mode (Behrend et al., 2008) can mitigate this problem. The results of our work presented in this paper have shown that it's possible to add more observations of southern sources without degradation of the tested geodetic products, such as EOP and baseline length repeatability. Indeed, more study is needed to find an optimal trade-off between the quality of geodetic and astrometric (CRF) products. More detailed investigations are anticipated for different R1, R4, and other IVS network configurations and an extended list of southern sources. In particular, inclusion of non-ICRF sources shall be considered at the next stage, as well as sources near the southern polar cap. Also, testing with VLI2010 recording parameters would be useful for future scheduling.

2.

3.

4.

5.

6.

T. Nilsson, B. Petrachenko, A. Rogers, G. Tuccari, and J. Wresnik. Recent Progress in the VLBI2010 Development. In: Observing our Changing Earth, IAG Symposia, Vol. 133, p. 833-840, 2008. doi: 10.1007/978-3-540-85426-5 96. ? ? J. Bohm, S. Bohm, T. Nilsson, A. Pany, L. Plank, H. Spicakova, K. Teke, and H. Schuh. The new Vienna VLBI Software VieVS. In: S. Kenyon, M. Pacino, and U. Marti (eds.), Proceedings of the IAG Scientific Assembly 2009, IAG Symposia, Vol. 136, p. 1007-1011, 2012. doi: 10.1007/978-3642-20338-1 126. J. Gipson. An introduction to Sked. In: D. Behrend and K. Baver (eds), Proceedings of the IVS 2010 General Meeting, NASA/CP-2010-215864, p. 77-84, 2010. C. Ma, E.F. Arias, G. Bianco, D.A. Boboltz, S.L. Bolotin, P. Charlot, G. Engelhardt, A.L. Fey, R.A. Gaume, A.-M. Gontier, R. Heinkelmann, C.S. Jacobs, S. Kurdubov, S.B. Lambert, Z.M. Malkin, A. Nothnagel, L. Petrov, E. Skurikhina, J.R. Sokolova, J. Souchay, O.J. Sovers, V. Tesmer, O.A. Titov, G. Wang, V.E. Zharov, C. Barache, S. Boeckmann, A. Collioud, J.M. Gipson, D. Gordon, S.O. Lytvyn, D.S. MacMillan, and R. Ojha. The Second Realization of the International Celestial Reference Frame by Very Long Baseline Interferometry. IERS Technical Note No. 35, A.L. Fey, D. Gordon, and ? C.S. Jacobs (eds.), Verlag des Bundesamts fur Kartographie ? und Geodasie, Frankfurt am Main, 2009. Z. Malkin, H. Schuh, C. Ma, and S. Lambert. Interaction between celestial and terrestrial reference frames and some considerations for the next VLBI-based ICRF. In: H. Schuh, ? S. Bohm, T. Nilsson, and N. Capitaine (eds.), Proceedings of the Journees 2011 Systemes de Reference Spatio-temporels, Ä ` ÄÄ Vienna, Austria, p. 66-69, 2012. ? J. Sun, A. Pany, T. Nilsson, J. Bohm, and H. Schuh. Status and future plans for the VieVS scheduling package. In: W. Alef, S. Bernhart, and A. Nothnagel (eds.), Proceedings of the 20th EVGA Meeting, Bonn, Germany, p. 44-48, 2011.

References
? 1. D. Behrend, J. Bohm, P. Charlot, T. Clark, B. Corey, J. Gipson, R. Haas, Y. Koyama, D. MacMillan, Z. Malkin, A. Niell,