1: %%
2: %% Beginning of file 'sample.tex'
3: %%
4: %% Modified 2005 December 5
5: %%
6: %% This is a sample manuscript marked up using the
7: %% AASTeX v5.x LaTeX 2e macros.
8:
9: %% The first piece of markup in an AASTeX v5.x document
10: %% is the \documentclass command. LaTeX will ignore
11: %% any data that comes before this command.
12:
13: %% The command below calls the preprint style
14: %% which will produce a one-column, single-spaced document.
15: %% Examples of commands for other substyles follow. Use
16: %% whichever is most appropriate for your purposes.
17: %%
18: \documentclass[12pt,preprint]{aastex}
19:
20: %% manuscript produces a one-column, double-spaced document:
21:
22: %%\documentclass[manuscript]{aastex}
23:
24: %% preprint2 produces a double-column, single-spaced document:
25:
26: %% \documentclass[preprint2]{aastex}
27:
28: %% Sometimes a paper's abstract is too long to fit on the
29: %% title page in preprint2 mode. When that is the case,
30: %% use the longabstract style option.
31:
32: %% \documentclass[preprint2,longabstract]{aastex}
33:
34: %% If you want to create your own macros, you can do so
35: %% using \newcommand. Your macros should appear before
36: %% the \begin{document} command.
37: %%
38: %% If you are submitting to a journal that translates manuscripts
39: %% into SGML, you need to follow certain guidelines when preparing
40: %% your macros. See the AASTeX v5.x Author Guide
41: %% for information.
42:
43: \newcommand{\vdag}{(v)^\dagger}
44: \newcommand{\myemail}{jwallin@gmu.edu}
45:
46: %% You can insert a short comment on the title page using the command below.
47: %%\slugcomment{Not to appear in Nonlearned J., 45.}
48: \slugcomment{Submitted to the \apj }
49: %% If you wish, you may supply running head information, although
50: %% this information may be modified by the editorial offices.
51: %% The left head contains a list of authors,
52: %% usually a maximum of three (otherwise use et al.). The right
53: %% head is a modified title of up to roughly 44 characters.
54: %% Running heads will not print in the manuscript style.
55:
56: \shorttitle{Testing Gravity in the Outer Solar System}
57: \shortauthors{Wallin et al.}
58:
59: %% This is the end of the preamble. Indicate the beginning of the
60: %% paper itself with \begin{document}.
61:
62:
63: \begin{document}
64:
65: %% LaTeX will automatically break titles if they run longer than
66: %% one line. However, you may use \\ to force a line break if
67: %% you desire.
68:
69:
70: \title{Testing Gravity in the Outer Solar System:
71: Results from Trans-Neptunian Objects}
72:
73: %% Use \author, \affil, and the \and command to format
74: %% author and affiliation information.
75: %% Note that \email has replaced the old \authoremail command
76: %% from AASTeX v4.0. You can use \email to mark an email address
77: %% anywhere in the paper, not just in the front matter.
78: %% As in the title, use \\ to force line breaks.
79:
80:
81: \author{John F. Wallin\altaffilmark{1}, David S. Dixon\altaffilmark{2}, \and Gary L. Page\altaffilmark{3} }
82:
83: \altaffiltext{1}{George Mason University, College of Science, Department of Computational and Data Sciences, Department of Physics and Astronomy, and Center for Earth Observing and Space Research (CEOSR), 4400 University Drive, MS 6A2, Fairfax, VA 22030; jwallin@gmu.edu.}
84: \altaffiltext{2}{Jornada Observatory, Las Cruces, NM; ddixon@cybermesa.com.}
85: \altaffiltext{3}{George Mason University, College of Science, Department of Computational and Data Sciences, 4400 University Drive, MS 6A2, Fairfax, VA 22030; gpage@gmu.edu.}
86:
87:
88:
89: %% Notice that each of these authors has alternate affiliations, which
90: %% are identified by the \altaffilmark after each name. Specify alternate
91: %% affiliation information with \altaffiltext, with one command per each
92: %% affiliation.
93:
94:
95: %% Mark off your abstract in the ``abstract'' environment. In the manuscript
96: %% style, abstract will output a Received/Accepted line after the
97: %% title and affiliation information. No date will appear since the author
98: %% does not have this information. The dates will be filled in by the
99: %% editorial office after submission.
100:
101: \begin{abstract}
102: The inverse square law of gravity is poorly probed by experimental
103: tests at distances of $\sim$ 10 AUs.
104: Recent analysis of the trajectory of the Pioneer 10 and 11
105: spacecraft have shown an unmodeled acceleration directed
106: toward the Sun which was not explained by
107: any obvious spacecraft systematics, and occurred when at
108: distances greater than 20 AUs from the Sun.
109: If this acceleration represents a departure from Newtonian gravity
110: or is indicative of an additional mass distribution in the outer solar system,
111: it should be detectable in the orbits of Trans-Neptunian Objects (TNOs).
112: To place limits on deviations from Newtonian gravity, we have selected a
113: well observed sample of TNOs
114: found orbiting between 20 and 100 AU from the Sun.
115: By examining their orbits with modified orbital fitting
116: software, we place tight limits on the perturbations
117: of gravity that could exist in this region of the
118: solar system.
119:
120: \end{abstract}
121:
122: %% Keywords should appear after the \end{abstract} command. The uncommented
123: %% example has been keyed in ApJ style. See the instructions to authors
124: %% for the journal to which you are submitting your paper to determine
125: %% what keyword punctuation is appropriate.
126:
127: \keywords{astrometry; celestial mechanics; ephemerides; interplanetary medium;
128: minor planets, asteroids; solar system: general}
129:
130: %% From the front matter, we move on to the body of the paper.
131: %% In the first two sections, notice the use of the natbib \citep
132: %% and \citet commands to identify citations. The citations are
133: %% tied to the reference list via symbolic KEYs. The KEY corresponds
134: %% to the KEY in the \bibitem in the reference list below. We have
135: %% chosen the first three characters of the first author's name plus
136: %% the last two numeral of the year of publication as our KEY for
137: %% each reference.
138:
139:
140: %% Authors who wish to have the most important objects in their paper
141: %% linked in the electronic edition to a data center may do so by tagging
142: %% their objects with \objectname{} or \object{}. Each macro takes the
143: %% object name as its required argument. The optional, square-bracket
144: %% argument should be used in cases where the data center identification
145: %% differs from what is to be printed in the paper. The text appearing
146: %% in curly braces is what will appear in print in the published paper.
147: %% If the object name is recognized by the data centers, it will be linked
148: %% in the electronic edition to the object data available at the data centers
149: %%
150: %% Note that for sources with brackets in their names, e.g. [WEG2004] 14h-090,
151: %% the brackets must be escaped with backslashes when used in the first
152: %% square-bracket argument, for instance, \object[\[WEG2004\] 14h-090]{90}).
153: %% Otherwise, LaTeX will issue an error.
154:
155: \section{Introduction}
156:
157: The theory of General Relativity (GR) has been
158: verified with a wide variety highly sensitive
159: of experiments. The effects of time dilation,
160: gravitational radiation (via timing of binary pulsars),
161: and gravitational lensing have been tested to very high precision.
162: However, most of the experiments that test GR are in the strong
163: limit of gravity, where the gravitational field and
164: associated mass density are typical for stars and compact
165: objects. Even in the solar system, we see the effects
166: of GR on the precession of Mercury's orbit as well as in other
167: precision experiments. However, in the weak limit of
168: gravity when objects are moving slowly, GR reduces to the familiar
169: Newtonian form of the inverse square law (\cite{2006LRR.....9....3W}).
170: This law is used in orbital dynamics to predict the location of planets
171: with objects more than $\sim$ 1 AU from the Sun without including
172: relativistic corrections.
173: Although Newtonian gravity's inverse square law shows excellent
174: agreement with observed data throughout on scales of a few AUs,
175: testing gravity in the outer
176: solar system at distances greater than 20 AUs has been difficult.
177:
178: Since objects (TNOs in particular) in the outer
179: part of the solar system cannot be observed with radar,
180: determining their orbits is done using optical astrometric
181: observations coupled with limited spacecraft observations.
182: The accuracy of these observations
183: and the relatively long time span needed to observe the
184: outer planets has led to some difficulties in matching
185: their orbits to Newtonian gravity. Even after the discovery
186: of Pluto, the anomalies in
187: Neptune's orbit were attributed to
188: a perturbing 10th planet, until this
189: issue was resolved with modern measurements
190: of planetary mass obtained from spacecraft (\cite{talmadge88}).
191: The limited astrometric accuracy, the long orbital period, and
192: relatively short time since most of these objects have been
193: accurately observed has led to uncertainty their orbits and precluded using
194: them for accurate tests of the inverse square law.
195:
196: Other tests of the weak limit of gravity at distances greater than
197: $\sim$ 10 AU have generally met with limited success. The flat
198: rotation curves of galaxies, for example, have been generally
199: interpreted as evidence of dark matter. However, we have not yet
200: directly detected dark matter particles by any observational or
201: experimental technique. This has led some to interpret the flat
202: rotation curves of galaxies as possible evidence that the Newtonian
203: approximation breaks down in the weak field limit. Instead of
204: invoking the existence of dark matter, Modified Newtonian Dynamics
205: (MOND) (\cite{milgrom83}) was developed to provide an alternative
206: explanation of the observed flat rotations curves. This theory has
207: had good success at modeling the rotation curves of many galaxies
208: based only on the distribution of the old stellar population.
209: \cite{TeVeS} has presented a Lorentz-covariant theory of gravity known
210: as TeVeS that yields MOND in its weak field limit. Although MOND, TeVeS and
211: other alternative theories of gravity have not been verified, the idea
212: of Newtonian gravity breaking down in its weak limit must be
213: considered as an alternative to dark matter to explain galaxy rotation
214: curves.
215:
216: The orbits of periodic comets in our solar system also have
217: shown deviations from Newtonian gravity. These deviations
218: have been characterized as non-gravitational
219: forces (\cite{kro04, mar69, mar73, mil99}) and are generally
220: attributed to out-gassing
221: of the comets as they approach the Sun. Each comet that shows
222: these deviations from Newtonian motion are fit to a set of
223: three acceleration parameters based on astrometric observations
224: of the orbit, not on physical models of cometary out-gassing.
225: Because of the parametric nature of the fitting process,
226: the orbits of long period comets cannot
227: confirm that Newtonian gravity is consistent with orbits in the
228: outer part of the solar system.
229:
230:
231:
232: One obvious way to measure gravity in the outer solar system is by
233: using the high accuracy tracking data of spacecraft leaving the solar
234: system. When the Pioneer 10 and 11 spacecraft were about 20 AU from
235: the Sun, their tracking data showed a systematic unmodeled
236: acceleration of $(8.74\pm 1.33) \times 10^{-8} $ cm s$^{-2}$ directed
237: toward the Sun. This acceleration appears at
238: between 10 and 20 AU, and
239: then remains constant outside of about 20AU.
240: The analysis of this data is detailed in
241: \cite{and98,and02a,and02b}. Obvious explanations such as interactions
242: with the solar wind, scattering of diffuse gas off a warm spacecraft,
243: and electromagnetic effects have been considered. Thus far, there are
244: no convincing physical phenomena that can cause this acceleration.
245: Although unmodeled spacecraft systematics are the most likely
246: explanation, it is possible that some new physical phenomenon may be
247: responsible for this effect. There are currently preliminary plans
248: to develop a spacecraft to investigate the Pioneer Anomaly
249: directly (\cite{PA-Collaboration}). Additionally, there is on-going
250: work to reanalyze the Pioneer spacecraft tracking data (c.f. \cite{Turyshev2006a},
251: \cite{Turyshev2006b}, \cite{Turyshev2005a}) and there is
252: considerable debate about the meaning of the original results. Even so, it is
253: certainly clear that separating the effects of spacecraft dynamics
254: from gravitational deceleration is a difficult task when one is trying
255: to measure small deviations from Newtonian gravity.
256:
257:
258: \cite{2003Icar..165..219W} have looked at the
259: orbits of Oort Cloud comets in order to independently examine the Pioneer
260: effect. If the Pioneer effect was affecting comets, the gravitational
261: binding energy would be higher and galactic tides could not
262: play the dominate role in making these objects observable.
263:
264: In this paper, we use Trans-Neptunian Objects (TNOs) to place limits
265: on deviations from Newtonian gravity in the outer parts of the solar system.
266: The use of planetary orbits to measure these
267: deviations from Keplerian orbits is not new and similar analyses
268: have been completed by other authors using astrometric data on the
269: major planets
270: (c.f. \cite{talmadge88}, \cite{hogg91}, \cite{sereno06}, \cite{iorio06})
271: In these papers, the authors either examine the residuals of
272: the orbital fit or project the orbital trajectories
273: forward in time and look for the expected deviation
274: between existing theories and observational uncertainty.
275: There are limitations on both of these approaches.
276:
277: In the work by Iorio (\cite{iorio06}), the authors project orbital
278: paths from a set of orbital elements forward with and without the
279: Pioneer effect and show that the differences in position are well
280: outside the astrometric uncertainty. As we have shown in our
281: previous paper (\cite{2006ApJ...642..606P}), one must consider how
282: both the old and new observed positions will change the derived
283: orbital elements rather than just looking for a shift between
284: prediction and observation positions. When new astrometric positions
285: for an orbiting object are obtained, a new fit to the orbital elements
286: is created. If an external force perturbs an orbit, the values of the
287: elements will slowly change as the fitting algorithm attempts to
288: integrate the new data into the orbital model. If these elements
289: change slowly enough, the addition of new astrometric data may produce
290: values that are within the uncertainty of the original values of the
291: orbital elements. Of course, the newly fitted orbital parameters will
292: provide a good approximation to the perturbed orbit, and the
293: perturbations will go undetected unless the residuals are examined
294: over the entire orbital history. Just projecting the resulting from a
295: set of static orbital elements forward in time does not address the
296: more subtle issues of orbital dynamics which could mask the detection
297: of orbital perturbations that we are attempting to detect.
298:
299:
300:
301: Although anomalous
302: accelerations of a sufficient magnitude would certainly
303: show up in the residuals of the orbital fit,
304: it is difficult to directly relate systematic
305: variation in the residual to an upper limit on any
306: perturbative acceleration.
307: As far as we know, no one has conducted the analysis of the uncertainty
308: in any perturbing acceleration added to Newtonian gravity
309: directly from planetary observations and related them
310: back directly to the expected systematic change in the
311: residuals of the orbital fit.
312: The ``detection by modeling''
313: method (\cite{hogg91}) is general more sensitive than looking for
314: systematic changes in the residuals. However, the very use
315: of a specific model for orbital perturbation
316: can limit the types of residuals that
317: are being detected. Thus far, most of the searches for
318: orbital perturbations have been looking
319: for a localized planet rather a radially dependent distribution
320: of matter or deviations from the inverse square law.
321:
322:
323:
324: In this paper, we take advantage of the large body
325: of astrometric data that has recently become available
326: on Trans-Neptunian Objects (TNOs).
327: Our approach is to use an ensemble of objects that have
328: been found in the outer solar system and whose observations are
329: archived in the Minor Planet Center Extended Computer Service (ECS).
330: Extending the modeling technique
331: of \cite{2006ApJ...642..606P}, we fit the orbits using a modified orbital
332: fitting program that allows a radially directed force of arbitrary
333: strength to be added to gravitational accelerations already
334: calculated by the program. For each object,
335: we calculate this anomalous acceleration
336: along with a statistically derived error
337: using the well documented Bootstrap technique (\cite{wall03}, \cite{efron}).
338:
339: Even though most of these TNOs have only been
340: recently identified, some of them have long observational arcs
341: because of the reanalysis of
342: archival images. The ensemble of observations used in
343: this paper covers a combined total
344: of 562 years of observations over 24 objects, making it a very
345: sensitive data set for examining gravity in the outer solar system.
346:
347: Beyond the results of this study, the methodology presented in this
348: paper can be extended to new objects discovered with future large
349: sky surveys such as Pan-STARRS and LSST. Using this technique,
350: strict limits on the deviations from Newtonian gravity can be
351: found constraining the solar system dark matter distribution
352: as well as other alternative theories of gravity such as
353: MOND or TeVeS.
354:
355:
356:
357:
358: \section{Methodology}
359:
360: \subsection{The Sample}
361:
362: To investigate possible gravitational perturbations to
363: the inverse square law in the outer solar system,
364: we formed a sample of objects from the ECS.
365: Our sample was selected based on three criteria:
366: \begin{enumerate}
367: \item The object must be observed at least 20 AU from the
368: Sun, where the Pioneer anomaly was detected by \cite{and02a}.
369: \item The object must have been observed over at least seven
370: oppositions at a heliocentric distance greater than 20 AU.
371: \item There must be at least forty archived observations
372: of the object.
373: \end{enumerate}
374:
375: The first constraint is imposed because
376: the Pioneer anomaly was first unambiguously detected
377: in the spacecraft tracking data when it was more
378: than 20 AU from the Sun. The last two constraints
379: were derived empirically. Our analysis has shown that
380: orbits with less than forty observations over at
381: least seven oppositions simply are not well enough constrained
382: to produce accurate values of the orbital elements including
383: the perturbing acceleration.
384: When additional objects are included, the large errors associated with
385: their fits make them extraneous to the final weighted average of the
386: results.
387:
388: Using the first criterion, we searched the 2006 May 1 Minor Planet Center's
389: ECS database of planetary orbits (MPCORB.DAT) and extracted an
390: initial sample of 31 objects from the 294,488 entries.
391: Observational data for each of these objects was then extracted
392: from the Minor Planet Center's observational archives (mpn.arc),
393: and preliminary orbits
394: were fit using the OrbFit Consortiums OrbFit (version 3.3) program.
395: Using these fitted orbits, we rejected an additional
396: seven objects as unsuitable for our analysis
397: because they failed the second and third criteria.
398: These resulting list of twenty-four objects and their orbital
399: characteristics are listed in Table \ref{tbl-1}.
400:
401:
402: The model we are fitting to these data is very simple, and is
403: applied separately to each object in our sample. We use the
404: bootstrap technique to estimate errors in our
405: fits, and then explore the results for systematic trends based
406: on position and orbital parameters. Finally, we combine the
407: results to place a limit on deviations from the inverse square
408: law in the outer solar system using the ensemble of data.
409:
410: \subsection{Orbital Fitting}
411:
412: To search for perturbations on Newtonian gravity, we used a modified
413: version of the OrbFit program that is used to fit orbits to
414: observations of asteroids. This code is well documented and is
415: widely used in the field. For our study, we add an
416: additional term to the gravitational acceleration from the Sun. The
417: effective acceleration of gravity from the Sun becomes:
418: \begin{eqnarray} \label{eqn-1}
419: g_{eff} = -\left\{ \frac{G M_\odot}{R^2} + \kappa \right\} \hat{r}
420: \end{eqnarray}
421:
422: where
423: \begin{eqnarray} \label{eqn-2}
424: \kappa = \left\{
425: \begin{array}{cc}
426: 0, &{\rm\ \ } R< 20 {\rm \ AU} \\
427: \epsilon, &{\rm\ \ } R \ge 20 {\rm \ AU} \\
428: \end{array}
429: \right.
430: \end{eqnarray}
431:
432: Where $\epsilon$ is an arbitrary parameter we fit to the observed data
433: for each object.
434:
435: Although we realize this model is not physically
436: realistic, we adopt is based on four considerations. First of all, the
437: model is consistent with what was seen in the Pioneer data and
438: other solar system constrains on the inverse square law. The
439: anomalous acceleration is constant after approximately 20 AU. The
440: Pioneer tracking data shows this anomaly turns on between 10 and 20 AU
441: from the Sun (c.f. \cite{and04}, Fig. 7). The particular form of the
442: transition is not well constrained by data. We also know that
443: the inverse square law is well characterized in the inner solar system,
444: and more poorly constrained in the outer solar system. The use
445: of a transition fits this behavior.
446: Second, the model is very simple with only one
447: free parameter. Since we fit this equation to each object separately
448: and then later examine its dependence on a set of orbital parameters,
449: we are making very minimal assumptions about the any anomalous
450: perturbation. Since most objects are found within a narrow range of
451: distances from the Sun, fitting each object separately allows us to
452: investigate an possible dependence of $\epsilon$ on heliocentric
453: distance. This is also true with variables such as ecliptic longitude
454: and orbital parameters such as eccentricity.
455: Third, only one object
456: in our sample ever goes inside the 20 AU cutoff. The inner and outer
457: orbital radii during the observational arc of each object in our
458: sample is presented in Table 1. Including a
459: more complicated transition would add unnecessary complexity to the
460: fit and add no significant knowledge to gravitational perturbation in
461: the outer solar system. Finally, the single
462: object (42355) that does go inside 20 AU has a large error on the
463: final fit of $\epsilon$, and does not significantly bias our final
464: results. Thus, the final fit we are using is effectively
465: gravitational acceleration plus a fitted radially directed
466: acceleration. The representation of the transition region at 20 AU
467: has no significant effect in our conclusions.
468:
469:
470: Because we have introduced the new parameter $\epsilon$ into the code,
471: we converge on the best value of this perturbing acceleration by using a
472: modified bisection method to find value of $\epsilon$ that produces
473: the minimal residual. Although
474: there may be exceptions, brute force examination of the residual on
475: selected test cases has shown that there is a single value for this
476: minimum residual, and the values smoothly decrease toward this
477: minimum. Using this bisection method,
478: we are able to converge to a value
479: for the perturbing force at a suitable accuracy with only about twenty
480: iterations. The best fit is estimated by fitting a parabola to
481: the three points nearest the minimum
482: and interpolating the location of the minimum.
483: For any given set of observations, we calculate the value of $\epsilon$
484: along with the minimal residual.
485:
486: It is important to note that adding an extra parameter to any
487: model will inevitably lead to non-zero values in that parameter
488: in poorly characterized data. As we discuss in the next section,
489: it is critical to be able to
490: characterize the quality of the data before making conclusions
491: about the overall value of the parameter $\epsilon$.
492:
493: \subsection{Statistical Analysis and Reliability of the Results}
494: In order to have confidence in the results from this study,
495: we use the Bootstrap method to re-sample the observational files.
496: As described in \cite{efron}, there are two basic versions this
497: technique that can be applied to fits of data.
498:
499: The first method of `bootstrapping the observations'
500: directly resamples the observational file.
501: For a given observational file with $n$ entries, the bootstrap file samples the
502: table of observations $n$ times with replacement. The resulting data table
503: is of the same length as the original, but some entries
504: have been duplicated and others have been dropped. This method has limited utility
505: with orbital fitting, since the resampling can fundamentally change the character
506: of the fit. If, for example, the resampling drops a single critical observation from
507: 50 years ago, the overall quality and reliability of the fit will be substantially diminished.
508:
509: The second method of `bootstrapping the residuals'
510: initially fits the orbit using all the original observational
511: entries creating a model orbit. The residuals of the fit
512: are then resampled and added to the model orbit, creating
513: a set of synthetic observations. These synthetic observations
514: are created at the same time intervals as the original data.
515: For orbital data, this method is preferable since it doesn't
516: introduce the systematic bias that would occur from dropping and
517: duplicating observations.
518:
519: In both methods,
520: a new orbit is then fit to the synthetic observations.
521: The process is repeated, and the acceleration parameter $\epsilon$
522: is tabulated. A mean and standard deviation for $\epsilon$
523: is then derived from the ensemble of runs.
524: Details of this method and its statistical basis are discussed
525: elsewhere (\cite{efron}, \cite{wall03}). As Wall and Jenkins observe,
526: this technique, which seems to give something for nothing,
527: is well established. Additionally, the bootstrap method has been shown
528: to provide converging estimates to the underlying
529: statistical properties of the resampled data.
530:
531:
532: For our analysis using the bootstrap methods,
533: we created a set of 100 simulated orbits for each object.
534: When we bootstrapped the residuals of these orbits (method two
535: from above), the runs all converged and gave us an estimate of the
536: anomalous acceleration $\epsilon$.
537:
538: Bootstrapping the observations (method one from above) was
539: more problematic.
540: For some objects, some of the synthetic orbits
541: failed to converge because of the nature of the resampled
542: observations. In some cases, entire years of observational measurements
543: can be dropped because of the resampling being done in the method. At the
544: same time, duplicate observations are created giving extra weight
545: to arbitrary entries. The resampling inherent with
546: directly bootstrapping the observations can lead to
547: large gaps in the observational
548: arc that make the trial data sets fall below the criteria
549: of seven oppositions with forty observations we set for
550: our sample selection. Although we do not
551: reject these runs a priori, the results can be a failure
552: to find a robust orbit that fully converges .
553: Nevertheless, about 75\% of the runs that bootstrapped
554: the observations did converge in our analysis.
555:
556:
557: Although we believe this lack of convergence in some of our runs will not
558: likely lead to a significant bias in our results, we present
559: the results from bootstrapping the observations only for completeness.
560: As discussed above, the objects that have the highest fraction
561: of non-converging runs were those that have the shortest
562: and poorest sample of observational arcs. It is likely that
563: the values of the anomalous acceleration ($\epsilon$) and the
564: errors ($\delta \epsilon$) are being under-estimated
565: on these objects.
566: All eight of the objects that had convergence
567: rates of less than 90\% has error estimates of greater
568: than 100 times the Pioneer effect.
569: Since the best sampled orbits have errors so much smaller than those
570: that only marginally fit our criteria, the impact on our
571: final results is small. We further discuss the impact of the non-converging
572: runs in Section \ref{results}.
573:
574:
575: It is important to reiterate that the convergence problems were not present when
576: we bootstrapped the residuals. Since the time intervals and data were
577: much more consistent with the original data fits, the fitting process
578: was much more robust. For parametric fits, bootstrapping the residuals
579: is generally preferred over bootstrapping the observations
580: because of these issues (\cite{efron}). Although we present the results
581: from both methods, we believe the results from the bootstrapped residuals
582: are more reliable.
583:
584:
585:
586:
587: %This convergence problem occurs for this program
588: %even when the additional force term is not included, and can be corrected
589: %to some level by manually selecting which points will be used for
590: %the initial orbital fit.
591: % We attempted to automatically select a
592: %``good'' set of entries from the observation table to obtain a good fit.
593: %Although this improved the success rate of convergence of our runs, for
594: %some objects, a number of runs were still rejected.
595: %Because so many runs were required
596: %to obtain a statistically signficant results, we simply
597: %were not able to manually select the observations for each run
598: %and guarantee convergence for every simulation generated.
599:
600:
601: % The second method used to examine the robustness of our results was a
602: % sensitivity analysis on the input file. After reading in the initial
603: % observation files, we add a one-arcsecond Gaussian random perturbation
604: % for each entry in the file. The perturbed file is then run through
605: % the modified OrbFit program, and the best fit is found for the
606: % perturbing force. The results are tabulated for the run, and the
607: % process is repeated until 500 good runs are accumulated in the log files.
608: % As in the bootstrap method, not all runs converged in OrbFit. Additional
609: % runs were beyond the required 500 were needed to make up for these
610: % bad fits.
611:
612: \section{Results and Analysis \label{results}}
613:
614:
615: The results of the analysis using the bootstrapped residuals method
616: are shown in Table \ref{tbl-2}.
617: The second column in this table represents
618: the average acceleration ($\epsilon$ in equation \ref{eqn-1}) and
619: its standard deviation from the 100 run ensemble.
620: We also calculate the average residual and standard deviation
621: of the residual for the ensemble of runs.
622: In the `acceleration' column, we present the fitted anomalous
623: acceleration using all the data in the observational files
624: without using bootstrapping. We refer to this fit as our ``best fit''
625: model. No formal error can be found on for this ``best fit''
626: model, since they are derived from a single set of observations.
627: The last column represents the residual
628: we found by fixing $\epsilon=0$, thus providing no perturbative force.
629:
630: All the forces in the table and the text below
631: are measured in terms of the Pioneer
632: effect acceleration of $8.74 \times 10^{-8}$ cm s$^{-2}$. An acceleration of
633: one in these units would be expected if the Pioneer Anomaly was affecting the
634: orbits of these objects, while an acceleration of zero would
635: indicated consistency with standard Newtonian gravity.
636:
637: To examine the consistency of our results, in Figure \ref{fig-1}
638: we plot the measured error in the anomalous acceleration ($\delta \epsilon$)
639: against the absolute value of the anomalous acceleration
640: ($\epsilon$) for each
641: object in our sample. In this plot,
642: there is an obvious correlation between the error and the value of the
643: anomalous acceleration derive from our fits.
644: The objects with large accelerations are those with
645: large errors in our sample, suggesting the true value of the
646: acceleration $\epsilon$ should be near zero.
647:
648: Figure \ref{fig-2} shows the relationship between the measured error
649: in the anomalous acceleration and the observed observational
650: arc in radians. As expected, the best characterized objects
651: have longer observational arcs. However, other factors such as
652: frequency of sampling also play a critical role in reducing the
653: errors in orbital determination. Regular observations over
654: a long time period are likely to yield significant improvements
655: of these results.
656:
657: In both Figure \ref{fig-1} and Figure \ref{fig-2}, the results are what we
658: would have expected in this experiment. Specifically, the estimates
659: we make of our errors seem to be consistent with the behavior of the
660: estimates of the accelerations. Based on the results, we believe error
661: estimates we derive using the bootstrap technique provides us with
662: a reliable measurement of the error associated with the fitted values
663: of $\epsilon$ on each
664: object. With an estimate of the
665: errors on $\epsilon$, we can calculate the weighted average acceleration
666: using the inverse of the variance as the weights (\cite{bevington}).
667: By doing this, we are assuming fitting
668: $\epsilon$ to a single value for all the objects. We will examine the
669: validity of this assumption below.
670:
671:
672: The weighted average for the ensemble of bootstrapped residual runs was
673: $0.10 \pm 0.16$ times the Pioneer acceleration where the
674: uncertainty is a one-sigma error. If we use
675: the bootstrap errors for the weighting and use the ``best fit''
676: models as the values for the run, the average acceleration
677: is identical to two significant digits.
678: These results are consistent with standard Newtonian gravity and a value of zero
679: for $\epsilon$. Using the `bootstrapped observation' method, we find the ensemble of bootstrap runs
680: gives us an average acceleration of $-0.23 \pm 0.28$ times the Pioneer acceleration.
681: If we use
682: the bootstrap errors for the weighting and use the ``best fit''
683: models as the values for the run, the average acceleration
684: is $0.03 \pm 0.28$. When we only include the objects that have had no problems with
685: convergence in any of their runs, the final results we obtain
686: for the acceleration is identical to three significant digits
687: as the weighted average of the overall sample. The lack
688: of convergence in the some runs using the bootstrapped observation method
689: creates no significant bias in our results.
690:
691:
692: To further examine gravity in the outer solar system, we looked at how
693: the measured acceleration is correlated with the position and orbital
694: parameters of the objects in our sample. Figure \ref{fig-3} shows the
695: relationship between the measured anomalous acceleration
696: ($\epsilon$) and the derived heliocentric distance of the orbits of
697: our objects. For clarity, only the objects with accelerations errors
698: less than ten times the Pioneer Anomaly were included in the plot.
699: There is no clear correlation between the heliocentric distance and
700: the anomalous acceleration. However, the error bars on our
701: measurements are too large to completely rule out such a correlation.
702:
703: In Figure \ref{fig-4}, we compare the anomalous accelerations for these TNOs
704: with the eccentricity of the orbit,
705: semi-major axis of the orbit, average heliocentric distance
706: during the observations, and ecliptic longitude.
707: No statistically significant correlations
708: were found in the data between mean acceleration $\epsilon$ and any of these
709: parameters. There is a weak correlation seen between semimajor axis and
710: $\epsilon$ as well as orbital
711: eccentricity and $\epsilon$. Although this should be examined
712: further with future observations,
713: it is well within the noise and not considered significant.
714: Because of the large uncertainty in some of these
715: measurements, detecting small trends in the data is not possible.
716: With that caveat, there is no evidence to support any systematic deviations
717: from the inverse square law as a function of these variables.
718: In general, we find no correlation between the orbital elements
719: or positions of our objects and the
720: results we find for the anomalous acceleration.
721:
722:
723: \section{Conclusions}
724:
725:
726: In this paper, we have presented a new method using orbital
727: measurements of an ensemble of TNOs to measure deviations
728: from the inverse square law or gravity in the outer solar system. The method
729: relies on doing separate orbital fits for each object, and then
730: characterizing the accuracy of each fit using the bootstrap technique.
731: Since no significant systematic trends were detected in our
732: sample, we combined
733: the data from all the objects using a weighted average to place
734: limits on deviations from the gravitational inverse square law
735: in the outer solar system.
736: Using existing
737: data, we have measured the deviation from the inverse square law
738: to be $\delta a = 8.7 \times 10^{-9} \pm 1.6 \times 10^{-8}$
739: cm s$^{-2}$ directed outward from the Sun for objects at
740: heliocentric distances of 20 to 100 AUs.
741: This result is consistent with zero at the 1$\sigma$ limit.
742:
743:
744: Based on our analysis of the observational data of TNOs,
745: we find that the gravitational acceleration in the outer solar system
746: is inconsistent with the Pioneer anomaly
747: at the $\sim 5\sigma$ level using both variations of the bootstrap analysis.
748: All of our results
749: were consistent (within $1 \sigma$) with Newtonian gravity
750: without any additional radial perturbative forces. This suggests that the
751: deceleration seen in the Pioneer tracking data was probably the result
752: of spacecraft systematics rather than exotic physics.
753: Even so, we cannot rule out the possibility that exotic physics is
754: affecting the Pioneer spacecraft trajectories. Our work only shows
755: that the trajectory data from the Pioneer Spacecraft is inconsistent
756: to what we see in large, slowly moving rocks in the outer solar system
757: (\cite{2006ApJ...642..606P}).
758:
759: These results were derived using existing astrometric data on minor planets
760: in the outer part of the solar system. The use of the
761: bootstrap method in this analysis has allowed us to provide a test
762: of the reliability of each orbital fits. Although bootstrapping the observations
763: directly is not held to be as reliable as bootstrapping the residuals, both results
764: are consistent with our conclusions.
765: The combination of these two methods provide a cross check, albeit a weak one, on
766: our results.
767:
768: We found no evidence of any correlation
769: between the measured values of $\epsilon$ and the object's position in the
770: solar system or its orbital characteristics. However, we would have only detected such
771: a trend if it were strongly present in the data with an amplitude much greater
772: than the Pioneer effect acceleration.
773:
774: Overall, the results we present confirm the veracity of the Newtonian gravitational
775: potential in the outer parts of the solar system. However, future analysis
776: of astrometric data from Pan-STARRS and LSST will provide a much more sensitive test
777: of gravity. If
778: the number of TNOs in our sample is expanded by a factor of 100, and these
779: objects have a long arc of regularly observed positions, we will be able
780: to increase the sensitivity of our results by a factor of about ten, depending
781: on the arc lengths and rate of observations. The data sets from Pan-STARRS
782: and LSST that will be created within the next ten years
783: will provide strong limits on alternative gravitational theories such as MOND.
784:
785:
786: Additional future work will focus on how
787: well this technique works for finding accelerations using
788: ensembles of synthesized observations. Of specific interest is
789: the role that orbital eccentricity and the length and completeness
790: of the observational arc plays in the results we obtain from
791: fitting acceleration parameters to observational data.
792:
793:
794:
795:
796: \acknowledgments
797:
798: The authors wish to acknowledge the Minor Planet Center for
799: observational data, available through their Extended Computer
800: Service\footnote{\url{http://cfa-www.harvard.edu/iau/services/ECS.html}}.
801: Additionally, the excellent software packages developed and maintained
802: by the OrbFit consortium\footnote{\url{http://newton.dm.unipi.it/orbfit}}
803: allowed orbital calculations to be performed with the requisite
804: precision. The OrbFit program made use of JPL's DE405 ephemeris
805: data\footnote{\url{http://ssd.jpl.nasa.gov/eph\_info.html}} to
806: describe the dynamics of the solar system. Finally, the authors would
807: like to think Dr. James Gentle and Dr. Daniel Carr of George Mason University
808: for their helpful input regarding the statistical analysis of this data.
809:
810:
811:
812:
813: %% To help institutions obtain information on the effectiveness of their
814: %% telescopes, the AAS Journals has created a group of keywords for telescope
815: %% facilities. A common set of keywords will make these types of searches
816: %% significantly easier and more accurate. In addition, they will also be
817: %% useful in linking papers together which utilize the same telescopes
818: %% within the framework of the National Virtual Observatory.
819: %% See the AASTeX Web site at http://www.journals.uchicago.edu/AAS/AASTeX
820: %% for information on obtaining the facility keywords.
821:
822: %% After the acknowledgments section, use the following syntax and the
823: %% \facility{} macro to list the keywords of facilities used in the research
824: %% for the paper. Each keyword will be checked against the master list during
825: %% copy editing. Individual instruments or configurations can be provided
826: %% in parentheses, after the keyword, but they will not be verified.
827:
828: %%%%%{\it Facilities:} \facility{CXO}
829:
830: %% Appendix material should be preceded with a single \appendix command.
831: %% There should be a \section command for each appendix. Mark appendix
832: %% subsections with the same markup you use in the main body of the paper.
833:
834: %% Each Appendix (indicated with \section) will be lettered A, B, C, etc.
835: %% The equation counter will reset when it encounters the \appendix
836: %% command and will number appendix equations (A1), (A2), etc.
837:
838: %%%%\appendix
839:
840: %%%\section{Appendix material}
841:
842:
843: %% The reference list follows the main body and any appendices.
844: %% Use LaTeX's thebibliography environment to mark up your reference list.
845: %% Note \begin{thebibliography} is followed by an empty set of
846: %% curly braces. If you forget this, LaTeX will generate the error
847: %% "Perhaps a missing \item?".
848: %%
849: %% thebibliography produces citations in the text using \bibitem-\cite
850: %% cross-referencing. Each reference is preceded by a
851: %% \bibitem command that defines in curly braces the KEY that corresponds
852: %% to the KEY in the \cite commands (see the first section above).
853: %% Make sure that you provide a unique KEY for every \bibitem or else the
854: %% paper will not LaTeX. The square brackets should contain
855: %% the citation text that LaTeX will insert in
856: %% place of the \cite commands.
857:
858: %% We have used macros to produce journal name abbreviations.
859: %% AASTeX provides a number of these for the more frequently-cited journals.
860: %% See the Author Guide for a list of them.
861:
862: %% Note that the style of the \bibitem labels (in []) is slightly
863: %% different from previous examples. The natbib system solves a host
864: %% of citation expression problems, but it is necessary to clearly
865: %% delimit the year from the author name used in the citation.
866: %% See the natbib documentation for more details and options.
867:
868: \begin{thebibliography}{}
869:
870: \bibitem[Anderson, et al.(1998)]{and98} Anderson, John ~D., Laing, Philip ~A.,
871: Lau, Eunice ~L., Liu, Anthony ~S., Nieto, Michael Martin, \& Turyshev,
872: Slava ~G. 1998, \prl, 81:14, 2858
873: % Indication, from Pioneer 10/11, Galileo, and Ullysses Data, of an Apparent
874: % Anomalous, Weak Long-Range Acceleration
875:
876: \bibitem[Anderson, et al.(2002a)]{and02a} Anderson, John ~D., Laing, Philip ~A.,
877: Lau, Eunice ~L., Liu, Anthony ~S., Nieto, Michael Martin, \& Turyshev,
878: Slava ~G. 2002a, \prd, 65, 082004
879: % Study of the Anomalous Acceleration of Pioneer 10 and 11
880:
881: \bibitem[Anderson, et al.(2002b)]{and02b} Anderson, J. ~D., Turyshev, ~S., \&
882: Nieto, M. ~M. 2002b, \baas, 34, 1172
883:
884: \bibitem[Anderson, et al. (2004)]{and04} Anderson, J. ~D., Laing, Philip ~A., Lau, Eunice ~L.,
885: Liu, Anthony ~S., Nieto, Michael Martin, \& Turyshev,
886: Slava ~G. \prd, 65, 082004
887:
888: % Effect of the Pioneer Anomaly on Long-Period Comet Orbits
889:
890: \bibitem[Bekenstein(2004)]{TeVeS} Bekenstein, J. 2004, \prd, 70, 3509
891:
892: \bibitem[Bevington (1969)]{bevington} Bevington, P. R. 1969 {\it Data Reduction and Error Analysis of the Physical Sciences} (New York, NY: McGraw-Hill)
893:
894: % A Mission to Explore the Pioneer Anomaly
895: \bibitem[The Pioneer Explorer Collaboration: H.~Dittus et
896: al.(2005)]{PA-Collaboration} The Pioneer Explorer Collaboration:
897: H.~Dittus, et al.\ 2005, ArXiv General Relativity and Quantum Cosmology
898: e-prints, arXiv:gr-qc/0506139
899:
900:
901: % An Introduction to the Bootstrap
902: \bibitem[Efron \& Tibshirani (1988)]{efron} Efron, B. \& Tibshirani, R.J. 1998
903: {\it An Introduction to the Bootstrap} (Boca Rato, FL: CRC Press)
904:
905:
906: % Dynamical limits on dark mass in the outer solar system
907: \bibitem [Hogg, et al. (1991)]{hogg91} Hogg, D. W., Quinlan, G. D., \& Tremaine, S. 1991, \aj, 101(6), 2274
908: % title???
909:
910: % What do the orbital motions of the outer planets of the Solar System tell us about the Pioneer anomaly?
911: \bibitem[Iorio \& Giudice(2006)]{iorio06} Iorio, L., \&
912: Giudice, G.\ 2006, New Astronomy, 11, 600
913:
914:
915: \bibitem[Kr\'{o}likowska(2004)]{kro04} Kr\'{o}likowska, M. 2004, \aap, 427, 1117
916: % Long-Period Comets with Nongravitational Effects
917:
918: % Comets and Nongravitational Forces
919: \bibitem[Marsden(1969)]{mar69} Marsden, B. ~G. 1969, \aj, 74, 720
920:
921: % Comets and Nongravitational Forces, II
922: \bibitem[Marsden, Sekanina, \& Yeomans(1973)]{mar73} Marsden, B. ~G., Sekanina, Z., \& Yeomans,
923: D. ~K. 1973, \aj, 78, 211
924:
925: % Comets and Nongravitational Forces, V
926: \bibitem[Milani(1999)]{mil99}Milani, Andrea, 1999 \icarus, 137, 269
927:
928: % Original MOND paper
929: %A modification of the Newtonian dynamics as a
930: % possible alternative to the hidden mass hypothesis
931: \bibitem[Milgrom(1983)]{milgrom83} Milgrom, M., 1983, \apj, 270,365
932:
933:
934: \bibitem[Page et al.(2006)]{2006ApJ...642..606P} Page, G.~L., Dixon, D.~S.,
935: \& Wallin, J.~F.\ 2006, \apj, 642, 606
936:
937: % Dark matter vs. modifications of the gravitational
938: % inverse-square law. Results from planetary motion in
939: % the solar system
940: \bibitem[Sereno \& Jetzer(2006)]{sereno06} Sereno, M., \&
941: Jetzer, P.\ 2006, ArXiv Astrophysics e-prints, arXiv:astro-ph/0606197
942:
943: % paper probing the 5th force using solar system constraints
944: % Model-independent constraints on possible modifications of Newtonian gravity
945: \bibitem[Talmadge et al.(1988)]{talmadge88} Talmadge, C.,
946: Berthias, J.-P., Hellings, R.~W., \& Standish, E.~M.\ 1988, Physical Review
947: Letters, 61, 1159
948:
949: \bibitem[Turyshev et al.(2005)]{Turyshev2005a} Turyshev, S.~G.,
950: Nieto, M.~M., \& Anderson, J.~D.\ 2005, ArXiv General Relativity and
951: Quantum Cosmology e-prints, arXiv:gr-qc/0503021
952:
953: \bibitem[Turyshev, Nieto, \& Anderson(2005)]{tur05} Turyshev, S.~G.,
954: Nieto, M.~M., \& Anderson, J.~D. 2005, Am.J.Phys., 73, 1003
955: % Study of the Pioneer Anomaly: A Problem Set
956:
957: % a Study of the Pioneer Anomaly:. New Data and Objectives for New Investigation
958: \bibitem[Turyshev et al.(2006a)]{Turyshev2006a} Turyshev, S.~G., Toth,
959: V.~T., Kellogg, L.~R., Lau, E.~L., \& Lee, K.~J.\ 2006a, International
960: Journal of Modern Physics D, 15, 1
961:
962: % The Pioneer Anomaly and its Implications
963: \bibitem[Turyshev et al.(2006b)]{Turyshev2006b} Turyshev, S.~G.,
964: Nieto, M.~M., \& Anderson, J.~D.\ 2006b, EAS Publications Series, 20, 243
965:
966:
967: \bibitem[Wall \& Jenkins(2003)]{wall03} Wall, J.V., \& Jenkins, C.R. 2003 {\it Practical Statistics
968: for Astronomers} (Cambridge:UK, Cambridge University
969:
970: \bibitem[Whitmire \& Matese(2003)]{2003Icar..165..219W} Whitmire, D.~P., \&
971: Matese, J.~J.\ 2003, Icarus, 165, 219
972:
973: \bibitem[Will(2006)]{2006LRR.....9....3W} Will, C.~M.\ 2006, Living Reviews
974: in Relativity, 9, 3
975:
976:
977:
978:
979:
980:
981:
982:
983:
984:
985:
986: %
987: %
988: % \bibitem[Smullin et al.(2005)]{2005PhRvD..72l2001S} Smullin, S.~J., Geraci,
989: % A.~A., Weld, D.~M., Chiaverini, J., Holmes, S., \& Kapitulnik, A.\ 2005,
990: % \prd, 72, 122001
991: %
992: %
993: % \bibitem[Iorio(2005)]{2005gr.qc.....7041I} Iorio, L.\ 2005, ArXiv General
994: % Relativity and Quantum Cosmology e-prints, arXiv:gr-qc/0507041
995: %
996: %
997: % \bibitem[Adelberger(2004)]{2004APS..APR.A1001A} Adelberger, E.\ 2004, APS
998: % Meeting Abstracts, A1001
999: %
1000: %
1001: % \bibitem[Anderson et al.(2002)]{2002AAS...200.1206A} Anderson, J.~D., Lau,
1002: % E.~L., Turyshev, S., Williams, J.~G., \& Nieto, M.~M.\ 2002, Bulletin of
1003: % the American Astronomical Society, 34, 660
1004: %
1005: %
1006: % \bibitem[Kirillov \& Turaev(2002)]{2002PhLB..532..185K} Kirillov, A.~A., \&
1007: % Turaev, D.\ 2002, Physics Letters B, 532, 185
1008: %
1009: %
1010: % \bibitem[Kaspi et al.(1994)]{1994ApJ...428..713K} Kaspi, V.~M., Taylor,
1011: % J.~H., \& Ryba, M.~F.\ 1994, \apj, 428, 713
1012: %
1013: %
1014: % \bibitem[Talmadge et al.(1988)]{1988PhRvL..61.1159T} Talmadge, C.,
1015: % Berthias, J.-P., Hellings, R.~W., \& Standish, E.~M.\ 1988, Physical Review
1016: % Letters, 61, 1159
1017: %
1018: %
1019: % \bibitem[Felten(1984)]{1984ApJ...286....3F} Felten, J.~E.\ 1984, \apj, 286,
1020: % 3
1021: %
1022: %
1023: % \bibitem[Milgrom(1983)]{1983ApJ...270..365M} Milgrom, M.\ 1983, \apj, 270,
1024: % 365
1025:
1026:
1027:
1028:
1029:
1030:
1031:
1032: % Lessons Learned from the Pioneers 10/11 for a Mission to Test the Pioneer Anomaly
1033: % \bibitem[Turyshev et al.(2005)]{Turyshev2005b} Turyshev, S.~G.,
1034: % Nieto, M.~M., \& Anderson, J.~D.\ 2005, ArXiv Physics e-prints,
1035: % arXiv:physics/0502123
1036: %
1037: % \bibitem[Nye(2005)]{2005PlR....25e..12N} Nye, B.\ 2005, Planetary Report,
1038: % 25, 12
1039: %
1040: %% A Cosmological Explanation to the Pioneer Anomaly
1041: % \bibitem[Masreliez(2005)]{2005Ap&SS.299...83M} Masreliez, C.~J.\ 2005,
1042: % \apss, 299, 83
1043: %
1044: %
1045: %
1046: %% A Route to Understanding of the Pioneer Anomaly
1047: % \bibitem[Turyshev et al.(2004)]{2004gr.qc.....9117T} Turyshev, S.~G.,
1048: % Nieto, M.~M., \& Anderson, J.~D.\ 2004, ArXiv General Relativity and
1049: % Quantum Cosmology e-prints, arXiv:gr-qc/0409117
1050: %
1051: %
1052: %
1053: %
1054: % \bibitem[Savenko et al.(2004)]{2004cosp.meet.4575S} Savenko, Y.~V.,
1055: % Demyanenko, P.~O., \& Zinkovskiy, Y.~F.\ 2004, 35th COSPAR Scientific
1056: % Assembly, 4575
1057: %
1058: %
1059: % \bibitem[Anderson et al.(2002)]{2002AAS...201.4509A} Anderson, J.~D.,
1060: % Turyshev, S., \& Nieto, M.~M.\ 2002, Bulletin of the American Astronomical
1061: % Society, 34, 1172
1062:
1063:
1064:
1065:
1066:
1067:
1068:
1069:
1070:
1071: \end{thebibliography}
1072:
1073: \clearpage
1074:
1075: %% Use the figure environment and \plotone or \plottwo to include
1076: %% figures and captions in your electronic submission.
1077: %% To embed the sample graphics in
1078: %% the file, uncomment the \plotone, \plottwo, and
1079: %% \includegraphics commands
1080: %%
1081: %% If you need a layout that cannot be achieved with \plotone or
1082: %% \plottwo, you can invoke the graphicx package directly with the
1083: %% \includegraphics command or use \plotfiddle. For more information,
1084: %% please see the tutorial on "Using Electronic Art with AASTeX" in the
1085: %% documentation section at the AASTeX Web site,
1086: %% http://www.journals.uchicago.edu/AAS/AASTeX.
1087: %%
1088: %% The examples below also include sample markup for submission of
1089: %% supplemental electronic materials. As always, be sure to check
1090: %% the instructions to authors for the journal you are submitting to
1091: %% for specific submissions guidelines as they vary from
1092: %% journal to journal.
1093:
1094: %% This example uses \plotone to include an EPS file scaled to
1095: %% 80% of its natural size with \epsscale. Its caption
1096: %% has been written to indicate that additional figure parts will be
1097: %% available in the electronic journal.
1098:
1099: %\begin{figure}
1100: %\epsscale{1.0}
1101: %\plotone{fig1.eps}
1102: %\caption{Summary of non-gravitational acceleration from TNO orbital analysis.\label{fig-1}
1103: %}
1104: %\end{figure}
1105:
1106: \begin{figure}
1107: \epsscale{1.0}
1108: \plotone{f1.eps}
1109: \caption{ Error of the measured anomalous acceleration $\Delta \epsilon$ vs the magnitude of measured
1110: anomalous acceleration $|\epsilon|$.
1111: The error of the measured acceleration is correlated with the strength of the acceleration, indicating
1112: the true value of the acceleration is small. All accelerations are measured
1113: in units of the measured Pioneer Anomaly ($8.74 \times 10^{-8} $ cm s$^{-2}$).
1114: Errors were derived using one standard deviation from one hundred
1115: trial runs of the bootstrap analysis for each object.
1116: \label{fig-1}
1117: }
1118: \end{figure}
1119:
1120:
1121: \begin{figure}
1122: \epsscale{1.0}
1123: \plotone{f2.eps}
1124: \caption{ Error of the measured anomalous acceleration ($\Delta
1125: \epsilon$) vs the observed geocentric arc of observations measured in
1126: radians. Objects with very short observational arcs ($< 0.2$) radians
1127: are poorly characterized and do not have errors small enough to
1128: contribute to the analysis. A long arc-length of observations by
1129: itself is not sufficient for small errors in the orbital fit.
1130: Additional factors such as the sampling rate and the quality of the
1131: astrometry play a role the reliability of the orbital fit as well.
1132: All accelerations are plotted in units of the measured Pioneer Anomaly
1133: ($8.74\times 10^{-8} $ cm s$^{-2}$). The acceleration errors $\Delta
1134: \epsilon$ were derived using one standard deviation from one hundred
1135: runs of the bootstrap analysis for each object.
1136: \label{fig-2}}
1137: \end{figure}
1138:
1139:
1140:
1141: \begin{figure}
1142: \epsscale{1.0}
1143: \plotone{f3.eps}
1144: \caption{ Anomalous acceleration $\epsilon$ vs average distance of the object from the Sun in AU over
1145: the observed arc.
1146: There is no apparent trend between the measured anomalous acceleration and the distance from the
1147: Sun. Only the objects with errors less than $\pm 10$ times the acceleration of the
1148: Pioneer effect were included in this plot for clarity. All accelerations are plotted
1149: in terms of the measured Pioneer Anomaly ($8.74\times 10^{-8} $ cm s$^{-2}$).
1150: The acceleration errors ($\Delta \epsilon$) were derived using one standard deviation from one hundred
1151: trial runs of the bootstrap analysis for each object.
1152: \label{fig-3}
1153: }
1154: \end{figure}
1155:
1156:
1157: \begin{figure}
1158: \epsscale{1.0}
1159: \plotone{f4.eps}
1160: \caption{ The measured anomalous acceleration ($\epsilon$) vs orbital inclination,
1161: ecliptic longitude, semimajor axis, and orbital eccentricity.
1162: Only the objects with errors less than $\pm 10$ times the acceleration of the
1163: Pioneer effect were included in this plot for clarity.
1164: There is no evidence of any statistically significant
1165: correlation between these four parameters and
1166: the corresponding acceleration.
1167: All accelerations are plotted
1168: in terms of the measured Pioneer Anomaly ($8.74\times 10^{-8} $ cm s$^{-2}$).
1169: The acceleration errors ($\Delta \epsilon$) were derived using one standard deviation from one hundred
1170: trial runs of the Bootstrap analysis for each object.
1171: \label{fig-4}}
1172: \end{figure}
1173:
1174:
1175:
1176: \clearpage
1177:
1178: %% Here we use \plottwo to present two versions of the same figure,
1179: %% one in black and white for print the other in RGB color
1180: %% for online presentation. Note that the caption indicates
1181: %% that a color version of the figure will be available online.
1182: %%
1183:
1184: %\begin{figure}
1185: %\plottwo{f2.eps}{f2_color.eps}
1186: % \caption{A panel taken from Figure 2 of \citet{rudnick03}.
1187: % See the electronic edition of the Journal for a color version
1188: % of this figure.\label{fig2}}
1189: % \end{figure}
1190: %
1191: % %% This figure uses \includegraphics to scale and rotate the still frame
1192: % %% for an mpeg animation.
1193: %
1194: % \begin{figure}
1195: % %\includegraphics[angle=90,scale=.50]{f3.eps}
1196: % \caption{Animation still frame taken from \citet{kim03}.
1197: % This figure is also available as an mpeg
1198: % animation in the electronic edition of the
1199: % {\it Astrophysical Journal}.}
1200: % \end{figure}
1201: %
1202: %% If you are not including electonic art with your submission, you may
1203: %% mark up your captions using the \figcaption command. See the
1204: %% User Guide for details.
1205: %%
1206: %% No more than seven \figcaption commands are allowed per page,
1207: %% so if you have more than seven captions, insert a \clearpage
1208: %% after every seventh one.
1209:
1210: %% Tables should be submitted one per page, so put a \clearpage before
1211: %% each one.
1212:
1213: %% Two options are available to the author for producing tables: the
1214: %% deluxetable environment provided by the AASTeX package or the LaTeX
1215: %% table environment. Use of deluxetable is preferred.
1216: %%
1217:
1218: %% Three table samples follow, two marked up in the deluxetable environment,
1219: %% one marked up as a LaTeX table.
1220:
1221: %% In this first example, note that the \tabletypesize{}
1222: %% command has been used to reduce the font size of the table.
1223: %% We also use the \rotate command to rotate the table to
1224: %% landscape orientation since it is very wide even at the
1225: %% reduced font size.
1226: %%
1227: %% Note also that the \label command needs to be placed
1228: %% inside the \tablecaption.
1229:
1230: %% This table also includes a table comment indicating that the full
1231: %% version will be available in machine-readable format in the electronic
1232: %% edition.
1233:
1234: \clearpage
1235:
1236:
1237:
1238: \begin{deluxetable}{cccccccc}
1239: \tabletypesize{\scriptsize}
1240: %\rotate
1241: \tablecaption{TNO Objects Investigated \label{tbl-1}}
1242: \tablewidth{0pt}
1243: \tablehead{
1244: \colhead{Name} & \colhead{\# Observations\tablenotemark{a}} & \colhead{a\tablenotemark{b}} (AU) &
1245: \colhead{e\tablenotemark{c}} & \colhead{R1\tablenotemark{d} (AU)}
1246: & \colhead{R2\tablenotemark{d} (AU)} &
1247: \colhead{Time Span\tablenotemark{e} (yr)} &\colhead{Oppositions}
1248: }
1249: \startdata
1250: \input{tab1.tex}
1251: \enddata
1252: %% Text for table notes should follow after the \enddata but before
1253: %% the \end{deluxetable}. Make sure there is at least one \tablenotemark
1254: %% in the table for each \tablenotetext.
1255: %\tablecomments{Table \ref{tbl-1} is published in its entirety in the
1256: %electronic edition of the {\it Astrophysical Journal}. A portion is
1257: %shown here for guidance regarding its form and content.}
1258: \tablenotetext{a}{Number of observations }
1259: \tablenotetext{b}{Semi-major axis (AU)}
1260: \tablenotetext{c}{Eccentricity}
1261: \tablenotetext{d}{Radius from the Sun at the beginning (R1) and end (R2) of the observations (AU)}
1262: \tablenotetext{e}{Time span between first and last observations (years)}
1263: \end{deluxetable}
1264:
1265:
1266: \clearpage
1267:
1268: % \begin{deluxetable}{rccccc}
1269: % \tabletypesize{\scriptsize}
1270: % %\rotate
1271: % \tablecaption{TNO Bootstrap Analysis \label{tbl-2}}
1272: % \tablewidth{0pt}
1273: % \tablehead{
1274: % \colhead{Object Name}
1275: % & \colhead{Bootstrap Acceleration\tablenotemark{a}}
1276: % & \colhead{Bootstrap Residual\tablenotemark{b} }
1277: % & \colhead{Acceleration\tablenotemark{c}}
1278: % & \colhead{Residual \tablenotemark{d}}
1279: % & \colhead{\# of Converged Runs \tablenotemark{e}}
1280: % }
1281: % \startdata
1282: % \input{tbl-2.tex}
1283: % \enddata
1284: % %Text for table notes should follow after the \enddata but before
1285: % %the \end{deluxetable}. Make sure there is at least one \tablenotemark
1286: % %in the table for each \tablenotetext.
1287: % %\tablecomments{Table \ref{tbl-1} is published in its entirety in the
1288: % %electronic edition of the {\it Astrophysical Journal}. A portion is
1289: % %shown here for guidance regarding its form and content.}
1290: % \tablenotetext{a}{The average of the best fit to the anomalous acceleration $\epsilon$ in terms of the
1291: % Pioneer Anomaly ($8.74\times 10^{-8}$ cm s$^{-1}$
1292: % along with the 1$\sigma$ error from the Bootstrap analysis.}
1293: % \tablenotetext{b}{Average residual and its 1$\sigma$ error from the Bootstrap analysis.}
1294: % \tablenotetext{c}{The best fit to the anomalous acceleration $\epsilon$ in terms of the
1295: % Pioneer Anomaly ($8.74\times 10^{-8}$ cm s$^{-1}$). No formal error is available
1296: % on this measurement since it was derived from a single fit.}
1297: % \tablenotetext{d}{Residual of the best fit.}
1298: % \tablenotetext{e}{Number of runs that converged in the Bootstrap analysis. A total of
1299: % 100 runs were attempted for each object.}
1300: % \end{deluxetable}
1301:
1302:
1303:
1304:
1305: \clearpage
1306:
1307: \begin{deluxetable}{rccccc}
1308: \tabletypesize{\scriptsize}
1309: %\rotate
1310: \tablecaption{TNO Bootstrap Analysis \label{tbl-2}}
1311: \tablewidth{0pt}
1312: \tablehead{
1313: \colhead{Object Name}
1314: & \colhead{Bootstrap Acceleration\tablenotemark{a}}
1315: & \colhead{Bootstrap Residual\tablenotemark{b} }
1316: & \colhead{Acceleration\tablenotemark{c}}
1317: & \colhead{Residual \tablenotemark{d}}
1318: %& \colhead{\# of Converged Runs \tablenotemark{e}}
1319: }
1320: \startdata
1321: \input{tab2.tex}
1322: \enddata
1323: %Text for table notes should follow after the \enddata but before
1324: %the \end{deluxetable}. Make sure there is at least one \tablenotemark
1325: %in the table for each \tablenotetext.
1326: %\tablecomments{Table \ref{tbl-1} is published in its entirety in the
1327: %electronic edition of the {\it Astrophysical Journal}. A portion is
1328: %shown here for guidance regarding its form and content.}
1329: \tablenotetext{a}{The average of the best fit to the anomalous acceleration $\epsilon$ in terms of the
1330: Pioneer Anomaly ($8.74\times 10^{-8}$ cm s$^{-1}$)
1331: along with the 1$\sigma$ error from the Bootstrap analysis.}
1332: \tablenotetext{b}{Average residual (arcseconds) and its 1$\sigma$ error from the bootstrap analysis.}
1333: \tablenotetext{c}{The best fit to the anomalous acceleration $\epsilon$ in terms of the
1334: Pioneer Anomaly ($8.74\times 10^{-8}$ cm s$^{-1}$). No formal error is available
1335: on this measurement since it was derived from a single fit.}
1336: \tablenotetext{d}{Residual (arcseconds) of the orbit assuming no perturbing force ($\epsilon = 0$).}
1337: %\tablenotetext{e}{Number of runs that converged in the Bootstrap analysis. A total of
1338: %100 runs were attempted for each object.}
1339: \end{deluxetable}
1340:
1341:
1342:
1343:
1344:
1345: \clearpage
1346:
1347:
1348:
1349:
1350: %% The optional \label should appear inside the \caption command.
1351:
1352:
1353: \end{document}
1354:
1355: