0708.3444/ms.tex
1: %\documentstyle[12pt,aasms4]{article}
2: \documentstyle[12pt,aaspp4,natbib]{article}
3: %\documentclass[12pt,preprint]{aastex}
4: 
5: \citestyle{aa}
6: 
7: \def\la{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}
8: \def\ga{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}
9: 
10: \slugcomment{Submitted to {\it The Astrophysical Journal}}
11: 
12: \received{2007 xxxxx xx}
13: \begin{document}
14: 
15: \title{PROMPT Observations of 
16: the Early-Time Optical Afterglow of GRB 060607A} 
17: 
18: \author{M. Nysewander\altaffilmark{1}\altaffilmark{2}, 
19: D. E. Reichart\altaffilmark{1}, 
20: J. A. Crain\altaffilmark{1},
21: A. Foster\altaffilmark{1},
22: J. Haislip\altaffilmark{1},
23: K. Ivarsen\altaffilmark{1},
24: A. Lacluyze\altaffilmark{1},
25: A. Trotter\altaffilmark{1}.
26: }
27: 
28: \altaffiltext{1}{Department of Physics and Astronomy, University of North
29: Carolina at Chapel Hill, Campus Box 3255, Chapel Hill, NC 27599}
30: \altaffiltext{2}{Space Telescope Science Institute, 3700 San Martin Drive,
31: Baltimore, MD, 21218;
32: mnysewan@stsci.edu}
33: 
34: \begin{abstract}
35: 
36: PROMPT (Panchromatic Robotic Optical Monitoring and Polarimetry Telescopes)
37: observed the early-time optical afterglow of GRB 060607A and obtained a 
38: densely sampled
39: multiwavelength light curve that begins only tens of seconds after the GRB.
40: Located at Cerro Tololo Inter-American Observatory in Chile, PROMPT is designed
41: to observe the afterglows of $\gamma$-ray bursts using multiple automated
42: 0.4-m telescopes
43: that image simultaneously in many filters when the afterglow is bright and
44: may be highly variable.
45: The data span the interval from 44 seconds 
46: after the GRB trigger to 3.3 hours in the $Bgri$ filters.  
47: We observe an initial 
48: peak in the light curve at 
49: approximately three minutes, followed by rebrightenings
50: peaking around 40 minutes and again at 66 minutes.  Although our data 
51: overlap with the early \textit{Swift} $\gamma$-ray and x-ray 
52: light curves, we do not see 
53: a correlation 
54: between the optical and high-energy flares.  
55: We do not find evidence for spectral evolution throughout the 
56: observations.
57: We model the variations in the light curves and find that the most likely
58: cause of the rebrightening episodes is a 
59: refreshment of the forward shock preceded by a rapidly fading
60: reverse shock component,
61: although other explanations are plausible.
62: 
63: \end{abstract}
64: 
65: \keywords{gamma rays: bursts}
66: 
67: \section{Introduction}
68: 
69: The general 
70: behavior of the broadband spectra and light curves of gamma-ray burst 
71: (GRB) afterglows has been shown to be well described by a jetted, 
72: relativistically expanding shell 
73: colliding with an external medium, which has been described as either that 
74: of a constant density 
75: \citep{spn98}
76: or wind-swept environment \citep{cl00}.  
77: Since the discovery of the first x-ray, optical and radio afterglows over
78: a decade ago \citep{cfh+97, vgg+97, fkn+97}
79: astronomers have successfully applied these models 
80: to derive the conditions of the shock and the 
81: properties of the surrounding medium.  
82: In the past decade, the number and quality of 
83: afterglow follow-up observations have improved dramatically.  The average 
84: response 
85: time for a GRB is now minutes instead of days, and 
86: for this rapid follow up, astronomers regularly use medium 
87: and large-aperture telescopes 
88: that can reach deep 
89: limiting magnitudes.  
90: 
91: However, in nearly all of the most densely sampled, high signal-to-noise 
92: light curves, significant 
93: variations are seen super-imposed upon the general behavior.
94: Three main physical 
95: scenarios 
96: have been proposed to explain these features: (1) the ``patchy 
97: shell" model \citep{kp00}, (2) delayed shocks or energy 
98: injections from a long-lived central engine \citep{rm98, kp200, sm00}, and
99: (3) variations in the circumburst density \citep{wl00, dl02}.
100: Although deviations from the simple power-law decay have been observed
101: since the beginning of the afterglow era (GRB 970508; \citealt{pmr98}),
102: the bright afterglows of GRB 021004 and GRB 030329 provide some of the most
103: densely sampled light curves, and both have been used as test cases for
104: these three mechanisms.
105: 
106: In the patchy shell model, random 
107: variations in the energy per unit angle in the 
108: outflow create the observed bumps in the light curve; the amplitude of these 
109: bumps is expected to decrease with time.
110: \citet{npg03} find that this model is slightly preferred over the
111: other two for GRB 021004 because of the morphology of the fluctuations 
112: and a consistent value for the electron index between low and high energies.
113: \citet{dcg+05} also find this to be an acceptable scenario in 
114: their detailed analysis
115: of the broadband (optical, near-infrared and millimeter) afterglow
116: light curve.
117: However, for GRB 030329
118: the patchy shell model has been ruled out
119: because the variations occurred after the jet break, when 
120: contributions from this effect ought to be negligible \citep{gnp03}.
121: 
122: Energy injections can be described as either delayed 
123: shocks that are hypothesized to be slower moving shells emitted by the 
124: central engine that catch up and impact the main shell of the forward
125: shock as it decelerates \citep{zm02}, or the result of a long-lived
126: central engine.
127: In this scenario, the light curve rises briefly as a result of the injection 
128: after which it resumes its decay with
129: an index similar to the previous index, but with a new normalization that
130: reflects the larger energy of the shock.  
131: Typically, the result of an
132: energy injection is an overall upward shift of the afterglow
133: light curve.
134: Additionally, when the impact occurs, 
135: it may also produce a bright reverse shock that 
136: propogates backwards through the ejecta in the comoving frame \citep{np03}.
137: This scenario was first proposed for
138: GRB 021004 by \citet{fyk+03} and GRB 030329 by \citet{gnp03}.
139: For GRB 021004, both \citet{bgj04} and 
140: \citet{dcg+05} find that the light curves, broadband 
141: spectral evolution and polarization signature of GRB 021004 can be 
142: well-modeled with a series of energy injections.  Due to the fact that 
143: the rebrightening episodes in the light curve of GRB 030329 
144: appear to be simple re-normalizations, 
145: both \citet{gnp03} and \citet{hcg06}
146: find refreshed shocks to be the most natural explanation of the event. 
147: 
148: Shocks resulting from the impact of the fireball upon 
149: density enhancements in the
150: surrounding medium can
151: cause significant rebrightenings in the GRB afterglow.
152: When the shockwave hits a higher density, the
153: flux sharply increases before it fades to match a light curve
154: that is described by the higher density, however, if the density jump is
155: large enough, it may also produce a reverse shock.
156: \citet{lrc+02} propose that the variations in the light curve of GRB 021004 are
157: likely due to moderate changes in density that only slightly modify the
158: dynamics of the fireball. 
159: Many groups find that density variations can not account for the
160: light curve of GRB 030329 due to the increase in flux normalization
161: after each episode \citep{gnp03, hcg06, uki+04}.
162: 
163: Early multi-color light curves of GRB afterglows are rare.  However, a number
164: of robotic telescopes have observed afterglows at times early enough to
165: compare the optical to the high-energy X-ray and $\gamma$-ray emission.
166: As pointed out by \citet{kp08}, this early-time emission
167: can be divided into three groups:  (1) The optical emission traces the
168: high-energy emission (e.g., GRB 041219A; Vestrand et al. 2006); the
169: optical emission does not trace the high-energy emission (e.g., GRB
170: 990123; Akerlof et al. 1999); and (3) the optical emission has two
171: components, one that traces the the high-energy emission and a smoother
172: component, probably the onset of the afterglow, that does not (e.g., GRB
173: 050820A; Vestrand et al. 2006).  GRB 0600607A falls into the second of
174: these categories (see \S4.1).
175: 
176: PROMPT \citep{rnm+05} has been designed specifically to observe the 
177: prompt optical and near-infrared 
178: emission from GRB afterglows simultaneously at multiple 
179: wavelengths.  The strength of this design is seen in the high-quality, 
180: densely sampled light curves of GRB 060607A presented in Figure~\ref{0607_lc}.  
181: Early in the afterglow's lifetime, it evolves on a rapid
182: time-scale, and 
183: only with \textit{simultaneous, multiwavelength} observations are 
184: we able to
185: properly characterize this phase.
186: With a data set of this quality we are 
187: able to ask detailed questions about the nature of the early afterglow 
188: and explore the possibility of chromatic variations.
189: In this paper, we present the general properties of the afterglow,
190: compare the optical, near-infrared and high-energy emission, and
191: focus on the interpretation of the variability
192: seen in the light curve.
193: In \S 2 we review the \textit{Swift} high-energy 
194: observations and present specific details of the 
195: PROMPT response.  In \S 3 we fit the standard afterglow and extinction
196: curve models and discuss the three distinct periods that occur within the first
197: two hundred minutes.  In \S 4 we compare the optical, near-infrared
198: and \textit{Swift} x-ray light curves and also discuss the nature
199: of the observed variations.  We draw conclusions in
200: \S 5.
201: 
202: \section{Observations}
203: 
204: The \textit{Swift} Burst Alert Telescope (BAT) discovered GRB 060607A
205: at 05:12:13 UT on June 7$^{th}$, 2006 \citep{zbg+06}.  The burst duration
206: is T$_{90}$ = 102 seconds (15 -- 350 keV) \citep{sbb+07} and exhibits a
207: triply-peaked structure \citep{tbb+06}.  Sixty-five seconds after the
208: initial trigger, the \textit{Swift}'s XRT found a fading x-ray afterglow 
209: \citep{pgb06},
210: and at 75 seconds, UVOT began to observe the bright optical afterglow
211: \citep{obz06}.  \citet{cdm+06} quickly reported a detection in
212: the near-infrared made by the REM telescope, which began 1.5 minutes after
213: the trigger \citep{mvm+06}.  \citet{nh06} reported 
214: $r$ magnitudes that detailed the rise of the early afterglow.  \citet{lvs+06},
215: began imaging with VLT 7.5 minutes after the BAT trigger and found an
216: afterglow redshift of z = 3.082.  
217: \citet{tbb+06} find that using the observed 15 -- 150 keV 
218: fluence of 2.6 $\pm$ $0.1 \times
219: 10^{-6}$ erg cm$^{-2}$, z = 3.082, 
220: and standard cosmology ($\Omega_{M}$ = 0.3,
221: $\Omega_{\Lambda}$ = 0.7, H$_{0}$ = 65), E$_{iso}$ = 1.1 x 10$^{53}$ ergs
222: in the rest frame 1 -- 1000 keV band.
223: 
224: Four 0.4-m PROMPT telescopes began observing the afterglow of GRB 060607A
225: in the Bessell $B$ and SDSS $gri$
226: filters
227: on June 7$^{th}$, 
228: 2006, 05:12:57 UT, 44 seconds after the initial satellite trigger
229: and 25 seconds after the GCN notification. 
230: Observations continued for 5.4 hours under the control of 
231: Skynet\footnote{Skynet is a dynamic, prioritized 
232: queue-scheduling system that controls a growing number of 
233: telescopes, currently spanning North and 
234: South America; http://skynet.unc.edu}, PROMPT's custom-designed
235: automation software.   
236: All images taken 
237: after 3.3 hours do not yield detections or limiting magnitudes deep or 
238: significant enough to be included in the analysis.  Table~\ref{0607_tab} 
239: presents the observations that are plotted in Figure~\ref{0607_lc}.
240: Figure~\ref{0607_lc} also includes 
241: the near-infrared $H$-band light curve measured by the
242: REM telescope \citep{mvm+06}. 
243: 
244: Zero, dark and flat-field calibration images were applied using IRAF's 
245: CCDPROC package, and, if necessary, images were combined to 
246: obtain better signal-to-noise.  Point-spread function photometry was applied 
247: via IRAF's DAOPHOT package to obtain final magnitudes.  
248: Zeropoints for each image were measured by reobserving the GRB field along with
249: observing photometric standards 
250: with PROMPT on the night of June 12, 2006.  Each PROMPT telescope 
251: uses a 1k x 1k Apogee Alta U47+ CCD, coated with either a midband 
252: or broadband coating in order to maximize the CCD response in the telescope's 
253: primary filter set. 
254: The cameras have fast-readout technology for fast cadence
255: imaging, with
256: an inter-exposure deadtime of only 2.5 seconds.
257: 
258: \section{Analysis}
259: 
260: Because of the
261: five densely sampled distinct light curves, we are able to see two 
262: significant
263: rebrightenings, although additional smaller fluctuations 
264: are suggested by the data.  
265: Therefore, we model the afterglow as the sum of three peaks, each consisting
266: of a smoothly broken, rising and falling power-law light curve and a simple
267: power-law spectrum, where
268: the spectrum is extinguished by both Milky-Way and source-frame dust and 
269: absorbed by hydrogen in the source frame and by the
270: Ly$\alpha$ forest along the line of sight:
271: \begin{equation}
272: F_{\nu}(t) = \sum_{n=1} ^3 e^{-\tau_{\nu}^{MW}} e^{-\tau_{\nu(1+z)}^{Ly \alpha}} e^{-\tau_{\nu(1+z)}^{source}} F_{n} \left(\frac{\nu}{\nu_o}\right)^{-\beta_{n}} \left[\left(\frac{t}{t_n}\right)^{-s_{n} \alpha_{n,1}} + \left(\frac{t}{t_n}\right)^{-s_{n} \alpha_{n,2}}\right]^{-1/s_{n}}.
273: \end{equation}
274: $\tau_{\nu}^{MW}$ is the Galactic extinction curve model of \citet{ccm89},
275: $\tau_{\nu(1+z)}^{Ly\alpha}$ is the Ly$\alpha$ forest
276: absorption model of \citet{r01},
277: and $\tau_{\nu(1+z)}^{source}$
278: is the source-frame extinction curve and Lyman-limit absorption model of
279: \citet{r01}
280: taking into account the Ly$\alpha$
281: dampening wing model of \citet{tkk+06}, which is a function of $N_H$.
282: $\alpha_{1n}$ and $\alpha_{2n}$ are the rising 
283: ($\alpha_{1n} > 0$) and falling ($\alpha_{2n} < 0$)
284: temporal
285: indices of the $n^{th}$ peak, $\beta_{n}$ is the spectral index, 
286: $t_{n}$ is the peak time,
287: $F_{n}$ is the normalization, $s_{n}$ is the smoothing
288: parameter, and $\nu_{o}$ is the
289: effective
290: frequency of the Sloan $r$ filter.
291: All magnitudes are converted to fluxes as prescribed by \citet{b79} and
292: \citet{bb88}.
293: Since the extinction and absorption models have features that are narrower 
294: than most photometric bands, we integrate equation (1) against the
295: appropriate filter transmissivity curve before fitting it to the data.
296: 
297: We fit this model to the data using Bayesian inference 
298: \citep[e.g.,][]{r01,ltv+01,grb+03,nrp+06}.
299: The posterior probability distribution is equal to the product
300: of the prior probability distribution and the likelihood function.  The
301: likelihood function is given by:
302: \begin{equation}
303: {\cal L} = \prod_{i=1}^N \frac{1}{\sqrt{2\pi(\sigma_i^2 + \sigma^2)}} \exp\left\{-\frac{1}{2}\frac{[y(\nu_i,t_i)- y_i]^2}{\sigma_i^2 + \sigma^2}\right\},
304: \end{equation}
305: where $N$ is the number of measurements, $y(\nu_{i}, t_{i})$ is the
306: integration of Equation 1 against the spectral curve of the $i$th
307: measurement at the time of the $i$th measurement; $y_{i}$ is the $i$th
308: measurement in units of log spectral flux; $\sigma_{i}$ is the
309: uncertainty in the $i$th measurement in the same units, and $\sigma$ is a
310: parameter, sometimes called the slop parameter, that models other,
311: small sources of uncertainty \citep{r01}.
312: We account for systematic error in zero point calibrations 
313: by allowing a magnitude offset for each filter where this offset is
314: constrained by 
315: a Gaussian prior of width given by the scatter in the 
316: relative magnitudes of the calibration stars.
317: These offsets 
318: are small and indicate calibration errors less than 0.1 magnitudes for the
319: PROMPT observations. 
320: The calibration offset adopted for the REM H-band light curve is $<$0.05
321: magnitudes.
322: 
323: Many of the parameters of the source-frame extinction curve model and all of
324: the parameters of the Ly$\alpha$ forest, Galactic extinction
325: curve, and Ly$\alpha$ damping wing models are constrained a 
326: priori.  The source-frame extinction curve
327: model of \citet{r01} is a function of eight parameters: the source-frame
328: V-band extinction magnitude $A_{V}$, $R_{V} = A_{V}/E(B-V)$, the intercept
329: $c_{1}$ and slope $c_{2}$ of the linear component of the source-frame UV
330: extinction curve, the strength $c_{3}$, width $\gamma$ and center $x_{o}$ of
331: the UV bump component of the extinction curve, and the strength $c_{4}$ of
332: the FUV excess component of the extinction curve.  The Ly$\alpha$ forest
333: absorption model of \citet{r01} is a function of a $D_{A}$, the flux
334: deficit.  
335: \citet{r01} determines prior probability distributions for $R_{V}$, $c_{1}$, \
336: $\gamma$, $x_{o}$, and $D_{A}$, which means that the values of these 
337: parameters can be weighted by fairly narrow distributions, the description
338: of which depends on other parameters ($c_{2}$ and $z$), a priori.  We adopt
339: these priors here.  The Galactic extinction curve model of \citet{ccm89}
340: is a function of $E(B-V)$ = 0.029 mag for this line
341: of sight \citep{sfd98} and a single parameter,
342: $R_{V}^{MW}$.  We adopt a prior for this parameter that is log
343: normally distributed with mean log 3.1 and width 0.1, which 
344: approximates the distribution of values of this parameter along
345: random lines of through the Galaxy.
346: The Ly$\alpha$ dampening wing model 
347: is a function of the density of neutral hydrogen, $N_{H}$, which
348: has a prior of $N_H = 7.9^{+3.8}_{-3.5} \times 10^{21}$ cm$^{-2}$
349: based on the measurement of \citet{pgb06}.
350: 
351: We fit an unconstrained model to the data and then impose relationships between 
352: the post-peak temporal and spectral indices of 
353: the first component --
354: these relationships depend upon the
355: local environment of the progenitor: a wind-swept (WIND, \citealt{cl00}) or
356: constant density (ISM, \citealt{spn98}) medium, and the placement of the cooling
357: break, $\nu_c$, above (BLUE) or below (RED) the observed frequencies.
358: For the ISM-RED and WIND-RED cases, $\alpha_{1,2} =
359: (3\beta_1
360: + 1)/2 = -(3p - 2)/4$; for the ISM-BLUE case, $\alpha_{1,2} = 3\beta_1/2 = -3(p
361: - 1)/4$; and
362: for the WIND-BLUE case, $\alpha_{1,2} = (3\beta_1 - 1)/2 = -(3p - 1)/4$, where
363: $p$ is the
364: power-law index of the electron-energy distribution.
365: In order
366: to test the spectral variability of the second and third components, only
367: $\beta_1$ and $\alpha_{1,2}$ are tied together using these relationships.
368: 
369: Using the PROMPT $Bgri$ light curves and the REM $H$-band
370: light curve \citep{mvm+06},
371: we find that WIND-BLUE with $p = 2.12^{+0.09}_{-0.07}$
372: is the most likely scenario, and is consistent with the general fit at the
373: 2.3 $\sigma$ confidence level.  
374: The ISM-BLUE and ISM/WIND-RED models are
375: ruled out at the 3.3 and 8.0 $\sigma$ confidence levels, respectively;
376: only the WIND-BLUE case is consistent with the observed shallow spectral index
377: $\beta = -0.70$.
378: Given the observed temporal index of $\alpha_{1,2} = -1.34^{+0.05}_{-0.07}$, for 
379: ISM-BLUE, $\beta_1 = -0.89^{+0.03}_{-0.05}$, 
380: and for
381: ISM/WIND-RED, $\beta_1 = -1.23 \pm$ 0.04.
382: Often, if a steep spectral slope is observed, degeneracy may exist between 
383: a steep intrinsic spectral index and the steepening effect of added
384: extinction, but for GRB 060607A, the observed shallow slope does not
385: allow for ambiguity.
386: 
387: Although we find a small amount of source-frame extinction, 
388: $A_{V} = 0.41^{+0.14}_{-0.30}$ mag ($A_{V} > 0$ 
389: at the 2.6 $\sigma$ credible level),
390: the parameters for the extinction curve model are not well
391: constrained.  
392: $c_{1}$ and $R_{V}$ depend upon $c_{2}$, and 
393: all are poorly constrained:
394: $c_{1} = -16.2^{+11.0}_{-7.3}$ and $c_{2} = 5.5^{+2.2}_{-3.3}$, while $R_{V}$
395: is completely unconstrained.
396: $c_{3}$ and $c_{4}$ are set to zero
397: because they are second order effects on 
398: the global shape of the spectrum.  
399: Similarly, although the shape of the light curve is well fit by the model
400: (see Figure~\ref{0607_lc}),
401: the exact fitting parameter values are again degenerate.  Because of the extreme
402: slopes involved with the second component, the parameters describing it
403: are not well-defined.  For the initial onset of the afterglow, 
404: $F_1 (\mu $Jy) $= 4.32^{+0.04}_{-0.14}$,
405: $\alpha_{1,1} = 2.46^{+0.23}_{-0.20}$, $\alpha_{1,2} = -1.34^{+0.05}_{-0.07}$, 
406: $\beta_1 =
407: -0.56^{+0.03}_{-0.05}$, log $t_1 $(days)$ = -2.70^{+0.01}_{-0.02}$, and $s_1 = 
408: 1.72^{+0.50}_{-0.38}$.
409: For the second component,
410: $F_2 (\mu $Jy) $ > 2.83$ (3 $\sigma$), $\alpha_{2,1} > 3.0$ (3 $\sigma$), $\alpha_{2,2} < -3.6$ 
411: (3 $\sigma$), $\beta_{2} = 
412: -0.49^{+0.15}_{-0.16}$, log $t_2 $(days)$ = -1.45^{+0.08}_{-0.15}$, and $s_2 <
413: 6.4$ (3 $\sigma$).
414: For the final component, 
415: $F_3 (\mu $Jy) $ = 2.53^{+0.20}_{-0.18}$, $\alpha_{3,1} > 0.57$ 
416: (2 $\sigma$), $\alpha_{3,2} = -1.28^{+0.16}_{-0.20}$,
417: $\beta_{3} =
418: -0.16^{+0.17}_{-0.24}$, log $t_3 $(days)$ = -1.36^{+0.04}_{-0.09}$, and $s_3 >
419: 3.29$ (1 $\sigma$).  
420: All values are cited with 1 $\sigma$ uncertainties, 
421: however, when only an upper or
422: lower limit could be placed, the value is cited with the most constraining
423: limit found (1, 2, or 3 $\sigma$).
424: Additionally log $N_{H} = 21.69^{+0.24}_{-0.21}$ cm$^{-2}$, 
425: and the slop parameter 
426: $\sigma = 0.07 \pm 0.01$ mag.
427: 
428: \section{Discussion}
429: 
430: Three peaks are evident in the prompt
431: $\gamma$-ray emission observed by the \textit{Swift} BAT: two
432: overlapping FRED profiles from t$_{0}$ -- 5 to t$_{0}$ + 40 seconds and a third
433: component at $\approx$100 seconds \citep{tbb+06}.  The 
434: PROMPT $g$ and $r$ light curves
435: overlap the third $\gamma$-ray peak, and 
436: although they are sparsely sampled at this
437: time, we do not observe any obvious corresponding features.  Similarly,
438: the peaks in the \textit{Swift}
439: XRT light curve around 90 and 250 seconds \citep{mvm+06}
440: do not appear in the PROMPT observations.  Furthermore, neither the initial 
441: afterglow peak
442: nor the later variations
443: at optical frequencies have corresponding features in the x-ray light curve,
444: although the x-ray light curve at the time of the peaks is not well-sampled.
445: A comparison of the PROMPT $r$ light curve and the \textit{Swift} XRT
446: x-ray light curve \citep{ebp+07} is presented in Figure~\ref{0607_xrt}.
447: Also, within the early optical and near-infrared light curves themselves,
448: there is no suggestion of chromatic variations.
449: 
450: \subsection{The Early Light Curve}
451: 
452: Both PROMPT and the REM telescope observed the early peak in the
453: afterglow at three minutes.
454: \citet{mvm+06} attribute this peak to the 
455: deceleration of the fireball
456: and in a detailed analysis, use this peak time
457: to calculate the Lorentz factor.
458: Given that the spectral index
459: should be $+$$1/3$ when $\nu_{opt} < \nu_m$
460: \citep[e.g.][]{gs02}, the passage of $\nu_m$, the synchrotron peak 
461: frequency, is ruled
462: out as a cause for the peak, and hence confirms the result of \citet{mvm+06}.  
463: Although the pre-peak
464: spectral index may show signs of evolution, the spectral index is
465: not positive (see Figure~\ref{0607_lc}), hence $\nu_m < \nu_{opt}$ before the onset
466: of observations.
467: 
468: The light curve at extreme early times does not appear to be well-fit by the
469: rising emission, but the discrepancies may be due to lower signal-to-noise of 
470: the early points.
471: In the very early light curve, deviations
472: from the model exist at the 1.8, 0.9 and 1.4$\sigma$
473: level for the initial $g$, $r$ and $i$ points respectively.
474: Early variations in the light curve can be interpreted as being due to activity
475: from the source, however, the early deviations do not
476: correlate with peaks in the X-ray light curve.  The lower time resolution
477: of the optical light curves makes it difficult to compare the low and 
478: high-energy emission.
479: 
480: Spectral evolution before the peak time is similarly difficult to quantify.
481: The spectral slope measured solely from these three points, centered at 
482: 84 seconds
483: is poorly constrained: $\beta = -6.5\pm3.3$.
484: The spectral slope measured from all points before the maximum light 
485: (during the time period from 65 to 174 seconds) is $\beta = 
486: -1.18\pm0.52$.
487: This slope is nearly consistent with the over-all slope of 
488: $\beta = -0.56^{+0.02}_{-0.05}$ for the first emission component.
489: 
490: \subsection{Comparison with X-Ray Afterglow}
491: 
492: The beginning of the \textit{Swift} era has brought many new questions about
493: the nature of x-ray afterglows.
494: X-ray light curves often do not correlate with their optical counterparts,
495: even when taking into account spectral breaks between the two frequencies.
496: Possible explanations for the disparity include
497: a long-lived central engine that continuously
498: refreshes the forward shock \citep{zfd+06},
499: two separate components to produce the x-ray and optical afterglows,
500: or variable microphysical parameters of the shock front \citep{pmb+06}.
501: Given $p = 2.12^{+0.09}_{-0.07}$ from the optical light curve,
502: we can examine the correlation between the high and low energy afterglows 
503: and determine if they
504: match predictions of the standard model scenarios.
505: 
506: The broadband optical to x-ray spectral flux distribution at twelve minutes
507: after the burst is presented in 
508: Figure~\ref{0607_spec}.  The x-ray afterglow flux is taken from \citet{mvm+06} 
509: and is plotted using the observed x-ray spectral index 
510: $\beta_x = -0.64 \pm$ 0.07.
511: $\beta_x$ was measured from the late-time PC data (which range from 600 seconds 
512: to two days),
513: and although the early
514: flares experienced spectral evolution \citep{pgb06}, we assume this value
515: of $\beta_x$ for the plot.
516: The SFD is plotted at twelve minutes, which is located immediately 
517: after the x-ray flares and 
518: near the break in the x-ray light curve when the temporal index shallows
519: from $\alpha_{x1} = -1.09 \pm$ 0.04 to $\alpha_{x2} = -0.45 \pm$ 0.03 
520: \citep{pgb06}.
521: The broadband spectral flux distribution 
522: indicates that, at this time, the low and high energy light curves may be
523: correlated.  
524: Although uncertain given the temporal scaling and insecure spectral slope, 
525: Figure~\ref{0607_spec} shows the 
526: extrapolation of the flux from the x-ray to optical band, which suggests that 
527: before the break in the x-ray light curve, one underlying mechanism may 
528: be producing both high and low-energy emission.
529: 
530: However, the best fit model to the optical and near-infrared afterglow 
531: is one with $\nu_c$ blueward of
532: optical frequencies ($\nu_c > \nu_{opt}$ at the 8.0 $\sigma$ credible level).  
533: Contrary to a constant density medium, for the synchrotron spectrum of a
534: shock expanding 
535: in a wind-swept environment, $\nu_c$
536: moves upwards through frequency space.
537: At twelve minutes, although $\nu_c > \nu_{opt}$, $\nu_c$ has not likely passed
538: above x-ray frequencies, hence the lack of a break in Figure~\ref{0607_spec}
539: would be inconsistent with our best fit.  
540: 
541: Indeed, the difference in temporal
542: slope at this time ($\alpha_{opt} = -1.34^{+0.05}_{-0.07}$ vs. $\alpha_{x1} =
543: -1.09 \pm 0.04$ or $\alpha_{x2} = -0.45 \pm 0.03$) 
544: indicates that there is a break between 
545: the high and low frequencies.
546: Therefore, it is possible that $\beta_x$ underwent a change in slope 
547: between twelve
548: minutes and when it was measured, and our assumed spectral slope plotted in
549: Figure~\ref{0607_spec} is incorrect.
550: In the WIND-BLUE fit: $p = 2.12^{+0.09}_{-0.07}$,
551: $\alpha_{opt} = -1.34^{+0.05}_{-0.07}$ and  
552: $\beta_{opt} = -0.56^{+0.03}_{-0.05}$.  If $\nu_{opt} < \nu_c < \nu_x$, 
553: then the x-ray data will be described by the WIND-RED case and 
554: the predicted values for the temporal and spectral x-ray slopes are:
555: $\alpha_{x, W-R} = -1.09^{+0.05}_{-0.07}$ and 
556: $\beta_{x,W-R} = -1.06^{+0.03}_{-0.05}$.
557: Note that the predicted value for the x-ray temporal slope matches
558: the observed early x-ray slope of $\alpha_{x, W-R} = \alpha_x = -1.09$, 
559: hence the pre-break ($t < 12$ minutes) period
560: of the x-ray light may be produced by the 
561: the same synchrotron component
562: as that of the optical light curve.
563: The break to the shallower, $\alpha_{x2} = -0.45 \pm 0.03$, slope 
564: would then correspond to the onset of a second phase of emission not
565: associated with the optical afterglow.
566: 
567: \subsection{Modeling the Variations}
568: 
569: Two clear achromatic
570: rebrightening episodes are observed in the optical and near-infrared
571: light curves at 40 and 66 minutes.
572: Here we consider possible scenarios for these peaks: the passage of a 
573: spectral break, the patchy-shell model, density variations, refreshed
574: shocks, and associated reverse shocks.
575: We do not expect that the variations in the light curve will be due to more
576: than one scenario, as it would be unlikely for multiple conditions
577: to be satisfied \citep{npg03}.
578: 
579: At the time near the peaks of the latter two components, the 
580: x-ray light curve is noisy
581: and not well-sampled, however it appears to flatten before
582: the beginning of the first optical rebrightening and
583: remains shallow until the end of the PROMPT light curve before dropping
584: suddenly at $\approx$4 hours (see Figure~\ref{0607_xrt}).  
585: Because $\nu_c$ is likely 
586: between the optical and x-ray bands,
587: the lack of features in the high-energy emission may suggest that the
588: episodes are due to density variations in the circumburst medium:
589: variations that depend upon changes in energy
590: would produce variations both above and below $\nu_c$.
591: However, without a clear understanding of the nature of the x-ray light
592: curve and the relationship between it and the optical, this claim is
593: questionable.
594: If, as suggested in Section \S4.1, the nearly flat phase of the 
595: x-ray emission is not produced by the same mechanism as that of the optical,
596: then evidence for variations above $\nu_c$ may be masked by the flux from 
597: a second
598: component.
599: 
600: \subsubsection{Spectral Evolution}
601: 
602: We first consider whether the peaks are due to the passage of a break --
603: although the indices are not generally well constrained,
604: they may suggest spectral evolution:
605: $\beta_1 = -0.56^{+0.03}_{-0.05}$, $\beta_2 = -0.49^{+0.15}_{-0.16}$ and 
606: $\beta_3 = -0.16^{+0.17}_{-0.24}$.
607: Deviations from our simple models of breaks in the electron distribution 
608: might influence the synchrotron spectrum and light curve in unexpected
609: ways. 
610: A model with no spectral evolution, where we impose that
611: $\beta_1 = \beta_3$ differs from the WIND-BLUE model at only the 1.9 $\sigma$ 
612: confidence level.
613: The observed change in spectral index ($\beta_3 = \beta_1 + 0.40 \pm 0.18$), 
614: from a moderate to 
615: shallow slope, is roughly consistent with the passage of 
616: $\nu_c$ from low to high
617: frequencies.  
618: However, the WIND and ISM-RED models are ruled out for the early data at the
619: 8.0 $\sigma$ confidence level, and hence we do not find this to be
620: a likely scenario.
621: 
622: \subsubsection{Patchy Shells}
623: 
624: The patchy-shell model, where the observed flux depends upon the
625: angular structure of the jet, has variations with a
626: timescale
627: longer than the observed time, $\Delta t > t$, and
628: more specifically, has decay times, $t_{decay} \approx t$ \citep{no04}.
629: At the end of the fluctuations, the light curve will fade to match the
630: extrapolation of the power-law light curve before the episode occurred.
631: The timescale for the first variation is short, $\Delta t \approx 27$
632: min, but the peak of the variation is at 33 min,
633: therefore the first relation is marginally satisfied.
634: However, the first peak decays rapidly, $t_{decay} \approx 15$ min, 
635: which is too extreme to
636: be expected from 
637: the timescale of the observations.
638: The second variation may satisfy both of these 
639: relationships;
640: however, the post-peak decay index does not indicate that 
641: the light curve is fading to its pre-episode flux.
642: Therefore, because the first variation is not consistent with the
643: predicted constraints and the decay of the second variation matches
644: the pre-episode index, the patchy shell model is
645: also unlikely.
646: 
647: \subsubsection{Density Variations}
648: 
649: Variations due to density enhancements in the circumburst environment  
650: can inflate the flux of the light curve that then settles into the 
651: afterglow solution describing the new environment. 
652: Generally, for a single density fluctuation, the light curve will asymptotically
653: relax to the 
654: pre-episode brightness.
655: However, if the shock impacts a shell after which the density increases
656: permanently,
657: then it is possible for the light curve to appear to renormalize.
658: In addition to this, as in the patchy-shell scenario, it is difficult for 
659: even a very sharp fluctuation in the density to produce
660: short time-scale fluctations 
661: in the light curve \citep{np03} with $\Delta t < t$.
662: Therefore, we do not find it likely 
663: that the first of the two variations is produced 
664: via this scenario.
665: 
666: However, if the jump in density is high enough ($\sim$21, \citealt{dl02}), then 
667: a reverse shock may form in addition to the enhancement of the forward shock. 
668: Due to the rapidly fading nature of the second feature, we find this to
669: be a plausible explanation for the second episode.
670: In this scenario, though, the apparent renormalization after the second
671: variation is a coincidence not predicted by the model, 
672: and relies upon conditions specific to the
673: circumburst density profile.
674: 
675: \subsubsection{Energy Injection}
676: 
677: If the bumps are due to two episodes of energy
678: injection, either by the continuing activity of the progenitor or 
679: by late internal shocks resulting from slower shells
680: impacting the
681: main forward shock,
682: then the fading temporal ($\alpha_{1,2} = \alpha_{2,2} =
683: \alpha_{3,2}$) and spectral ($\beta_1 = \beta_2 = \beta_3$) indices will 
684: remain nearly constant.
685: The energy injection will raise the flux of the light curve and
686: although the post-shock indices will reflect the new shock
687: conditions,
688: they should vary little from the previous state.
689: Due to the steepness of $\alpha_4$ this model is ruled out at the 
690: 7.5 $\sigma$ confidence level.
691: 
692: However, as in the case of density variations, it is possible that
693: internal shocks can cause a bright
694: reverse shock to form
695: and be observed
696: as a bright, rapidly fading flare in the GRB afterglow. 
697: In this case, the first and third variations will have nearly the same indices 
698: ($\alpha_{1,2} = \alpha_{3,2}$;
699: $\beta_1 = \beta_3$) and the second variation will fade rapidly 
700: ($\alpha_{2,2} \geq 2$).
701: Quantitatively, this model is consistent with the general model at the
702: 1.9 $\sigma$ uncertainty level.
703: Qualitatively, this is precisely what we see: the second episode at 
704: 66 minutes is a simple renormalization
705: after which the afterglows fades with the same initial slope, whereas the first
706: clearly fades more steeply. 
707: In fact, focusing on the pre- and post-episode slopes, a model 
708: with $\alpha_{1,2} = \alpha_{3,2}$ is consistent with a model
709: where the parameters are unconstrained at the 0.4 $\sigma$ confidence level.
710: Therefore, we find that a refreshed shock that was preceeded by a bright
711: reverse shock is the most likely cause of the two rebrightening episodes.
712: The observations are consistent with density variations, however, the
713: renormalization required in that circumstance would be merely 
714: coincidental, rather
715: than a prediction of the model.
716: 
717: \section{Conclusions}
718: 
719: We have presented the early-time PROMPT afterglow light curve for GRB 060607A,
720: fit a model to the data and found that the progenitor likely exploded
721: into a wind-swept medium and that the cooling frequency of the synchrotron
722: spectrum is above the optical frequencies before the beginning of the
723: PROMPT observations.
724: Observing correlations between peaks in
725: different energy ranges is a key tool for probing the underlying physics
726: of the central engine and its interaction with its environment.
727: Although the x-ray spectrum exhibited intense flares at early times,
728: these flares are not present in the optical light curves.
729: Likewise, optical flares at later times are not observed in
730: the x-ray light curve.  
731: In the PROMPT optical light curve we see an initial peak 
732: at 2.9 minutes, and later, 
733: two clear rebrightening episodes at 40 and 66 minutes that are not
734: observed at high energies.
735: We do not observe spectral evolution throughout the PROMPT dataset and 
736: constrain $\nu_m$ to be below and $\nu_c$ to be above optical 
737: frequencies by the onset of observations.
738: 
739: Many authors have cited the need for observations at both low and high energy
740: in order to distinguish between 
741: variations due to energy or density, and in this case the lack of features
742: in the high-energy light curve would indicate that the fluctations are due
743: to density.
744: However, without a clear understanding of the
745: mechanisms that produce the light curves, we cannot definitively relate
746: them.
747: We find that the most likely scenario for the observed fluctuations
748: is that of energy injection to the forward shock, either through a delayed 
749: shock or continuing activity of the central engine, which 
750: is strong enough to produce
751: a bright reverse shock that sweeps through the relativistic
752: material behind the main forward shock.
753: In the case of GRB 060607, although energy injection is the favored 
754: scenario,
755: the constraints imposed by the light curves are not conclusive proof.
756: 
757: As efforts to observe the afterglows of $\gamma$-ray bursts have 
758: strengthened, data
759: sets grow more robust.
760: Clearly, if variations are ubiquitous, the analysis of a poorly sampled
761: single light curve is difficult to trust.
762: The results obtained via 
763: modeling such a light curve 
764: that has undergone episodes of rebrightening will likely deviate from its true
765: behavior, and the steepness of
766: temporal decays
767: will be systematically underestimated.
768: PROMPT is a powerful tool that can produce densely sampled early-time
769: light curves when afterglows are bright and more 
770: likely to undergo rapid variations.
771: By observing simultaneously in multiple filters, we can test the early-time
772: afterglow for chromatic variations, measure spectral indices, and 
773: obtain multiple, independent
774: light curves.  Because of these factors, PROMPT will play a strong 
775: role in future afterglow follow-up 
776: studies.
777: 
778: \acknowledgements
779: DER very gratefully acknowledges support from NSF's MRI, CAREER, AAG, 
780: and PREST programs, NASA's APRA, Swift GI, and IDEAS programs, NC 
781: Space Grant's NIP program, and especially Leonard Goodman and Henry Cox. 
782: We gratefully acknowledge Don Smith for useful comments and collaboration
783: on the PROMPT project, and also T. Brennan,
784: M. Schubel and J. Styblova for their work as new members of the 
785: UNC GRB Follow-Up Group.
786: This work made use of data supplied by the UK Swift Science Data Centre at 
787: the University of Leicester.
788: 
789: \clearpage
790: 
791: \begin{deluxetable}{lcccc}
792: \tablecolumns{5}
793: \tablewidth{0pt}
794: \tablecaption{PROMPT Observations of the Afterglow of GRB 060607A}
795: \tablehead{\colhead{Mean Time (UT)} & \colhead{Exp. Time} & \colhead{Mean $\Delta$t (hr)} & \colhead{Filter} 
796: & \colhead{Magnitude}}
797: \startdata
798: Jun 7	5:13:38	&	15	s	&	0.0235	&	$B$	&	$>$16.61	\\
799: Jun 7	5:14:09	&	20	s	&	0.0322	&	$B$	&	16.44	$\pm$	0.10	\\
800: Jun 7	5:14:40	&	20	s	&	0.0408	&	$B$	&	15.69	$\pm$	0.07	\\
801: Jun 7	5:15:11	&	20	s	&	0.0494	&	$B$	&	15.47	$\pm$	0.06	\\
802: Jun 7	5:15:42	&	20	s	&	0.0581	&	$B$	&	15.63	$\pm$	0.07	\\
803: Jun 7	5:16:25	&	40	s	&	0.0700	&	$B$	&	15.59	$\pm$	0.04	\\
804: Jun 7	5:17:15	&	40	s	&	0.0839	&	$B$	&	15.84	$\pm$	0.05	\\
805: Jun 7	5:18:06	&	40	s	&	0.0981	&	$B$	&	16.03	$\pm$	0.06	\\
806: Jun 7	5:18:56	&	40	s	&	0.1119	&	$B$	&	16.21	$\pm$	0.07	\\
807: Jun 7	5:20:07	&	80	s	&	0.1317	&	$B$	&	16.51	$\pm$	0.09	\\
808: Jun 7	5:21:45	&	80	s	&	0.1589	&	$B$	&	16.55	$\pm$	0.05	\\
809: Jun 7	5:23:23	&	80	s	&	0.1861	&	$B$	&	16.86	$\pm$	0.06	\\
810: Jun 7	5:25:01	&	80	s	&	0.2133	&	$B$	&	17.16	$\pm$	0.11	\\
811: Jun 7	5:26:36	&	80	s	&	0.2397	&	$B$	&	17.25	$\pm$	0.12	\\
812: Jun 7	5:28:53	&	160	s	&	0.2778	&	$B$	&	17.60	$\pm$	0.09	\\
813: Jun 7	5:32:44	&	240	s	&	0.3420	&	$B$	&	17.77	$\pm$	0.09	\\
814: Jun 7	5:37:26	&	240	s	&	0.4203	&	$B$	&	17.89	$\pm$	0.08	\\
815: Jun 7	5:42:04	&	240	s	&	0.4974	&	$B$	&	17.85	$\pm$	0.07	\\
816: Jun 7	5:46:36	&	240	s	&	0.5730	&	$B$	&	17.75	$\pm$	0.08	\\
817: Jun 7	5:51:50	&	320	s	&	0.6601	&	$B$	&	17.98	$\pm$	0.07	\\
818: Jun 7	6:00:52	&	640	s	&	0.8110	&	$B$	&	18.56	$\pm$	0.09	\\
819: Jun 7	6:14:19	&	800	s	&	1.0350	&	$B$	&	18.61	$\pm$	0.11	\\
820: Jun 7	6:28:11	&	640	s	&	1.2662	&	$B$	&	18.91	$\pm$	0.11	\\
821: Jun 7	7:09:32	&	960	s	&	1.9553	&	$B$	&	19.53	$\pm$	0.19	\\
822: Jun 7	7:48:25	&	1200	s	&	2.6033	&	$B$	&	19.98	$\pm$	0.18	\\
823: Jun 7	8:18:06	&	1360	s	&	3.0981	&	$B$	&	19.75	$\pm$	0.14	\\
824: Jun 7	5:13:20	&	5	s	&	0.0185	&	$g$	&	16.63	$\pm$	0.15	\\
825: Jun 7	5:13:46	&	10	s	&	0.0258	&	$g$	&	16.11	$\pm$	0.10	\\
826: Jun 7	5:14:46	&	10	s	&	0.0425	&	$g$	&	15.04	$\pm$	0.03	\\
827: \enddata
828: \tablenotetext{a}{Upper limits are 3$\sigma$.}
829: \label{0607_tab}
830: \end{deluxetable}
831: 
832: \begin{deluxetable}{lcccc}
833: \tablecolumns{5}
834: \tablewidth{0pt}
835: \tablecaption{PROMPT Observations of the Afterglow of GRB 060607A (con't)}
836: \tablehead{\colhead{Mean Time (UT)} & \colhead{Exp. Time} & \colhead{Mean $\Delta$t (hr)} & \colhead{Filter}
837: & \colhead{Magnitude}}
838: \startdata
839: Jun 7   5:15:48 &       20      s       &       0.0597  &       $g$      &
840: 15.09   $\pm$   0.03    \\
841: Jun 7   5:17:17 &       20      s       &       0.0844  &       $g$      &
842: 15.37   $\pm$   0.02    \\
843: Jun 7	5:19:03	&	40	s	&	0.1139	&	$g$	&	15.82	$\pm$	0.02	\\
844: Jun 7	5:21:55	&	80	s	&	0.1617	&	$g$	&	16.10	$\pm$	0.01	\\
845: Jun 7	5:25:06	&	80	s	&	0.2147	&	$g$	&	16.67	$\pm$	0.03	\\
846: Jun 7	5:29:45	&	80	s	&	0.2922	&	$g$	&	17.20	$\pm$	0.03	\\
847: Jun 7	5:31:17	&	80	s	&	0.3178	&	$g$	&	17.30	$\pm$	0.05	\\
848: Jun 7	5:35:57	&	80	s	&	0.3956	&	$g$	&	17.38	$\pm$	0.04	\\
849: Jun 7	5:37:29	&	80	s	&	0.4211	&	$g$	&	17.31	$\pm$	0.04	\\
850: Jun 7	5:42:11	&	80	s	&	0.4994	&	$g$	&	17.27	$\pm$	0.04	\\
851: Jun 7	5:43:42	&	80	s	&	0.5247	&	$g$	&	17.36	$\pm$	0.04	\\
852: Jun 7	5:45:17	&	80	s	&	0.5511	&	$g$	&	17.33	$\pm$	0.04	\\
853: Jun 7	5:46:27	&	40	s	&	0.5706	&	$g$	&	17.37	$\pm$	0.07	\\
854: Jun 7	5:54:12	&	80	s	&	0.6997	&	$g$	&	18.04	$\pm$	0.06	\\
855: Jun 7	5:55:42	&	80	s	&	0.7247	&	$g$	&	18.04	$\pm$	0.07	\\
856: Jun 7	5:57:13	&	80	s	&	0.7500	&	$g$	&	18.05	$\pm$	0.05	\\
857: Jun 7	5:58:24	&	40	s	&	0.7697	&	$g$	&	18.15	$\pm$	0.07	\\
858: Jun 7	6:08:25	&	280	s	&	0.9367	&	$g$	&	18.52	$\pm$	0.17	\\
859: Jun 7	6:26:04	&	240	s	&	1.2307	&	$g$	&	18.41	$\pm$	0.08	\\
860: Jun 7	6:30:39	&	240	s	&	1.3071	&	$g$	&	18.41	$\pm$	0.05	\\
861: Jun 7	7:05:40	&	240	s	&	1.8907	&	$g$	&	19.20	$\pm$	0.17	\\
862: Jun 7	7:42:28	&	640	s	&	2.5042	&	$g$	&	19.48	$\pm$	0.06	\\
863: Jun 7	8:22:50	&	800	s	&	3.1768	&	$g$	&	19.74	$\pm$	0.08	\\
864: Jun 7	5:13:18	&	10	s	&	0.0181	&	$r$	&	15.98	$\pm$	0.31	\\
865: Jun 7	5:13:44	&	10	s	&	0.0253	&	$r$	&	15.75	$\pm$	0.16	\\
866: Jun 7	5:14:43	&	20	s	&	0.0417	&	$r$	&	14.34	$\pm$	0.03	\\
867: Jun 7	5:15:45	&	20	s	&	0.0589	&	$r$	&	14.32	$\pm$	0.03	\\
868: Jun 7	5:17:19	&	40	s	&	0.0850	&	$r$	&	14.68	$\pm$	0.02	\\
869: Jun 7	5:19:01	&	40	s	&	0.1133	&	$r$	&	15.11	$\pm$	0.02	\\
870: \enddata
871: \end{deluxetable}
872: 
873: \begin{deluxetable}{lcccc}
874: \tablecolumns{5}
875: \tablewidth{0pt}
876: \tablecaption{PROMPT Observations of the Afterglow of GRB 060607A (con't)}
877: \tablehead{\colhead{Mean Time (UT)} & \colhead{Exp. Time} & \colhead{Mean $\Delta$t (hr)} & \colhead{Filter}
878: & \colhead{Magnitude}}
879: \startdata
880: Jun 7   5:21:47 &       80      s       &       0.1594  &       $r$      &
881: 15.47   $\pm$   0.02    \\
882: Jun 7   5:25:03 &       80      s       &       0.2139  &       $r$      &
883: 16.04   $\pm$   0.03    \\
884: Jun 7   5:29:43 &       80      s       &       0.2917  &       $r$     &
885: 16.60   $\pm$   0.05    \\
886: Jun 7   5:31:12 &       80      s       &       0.3164  &       $r$      &
887: 16.57   $\pm$   0.05    \\
888: Jun 7	5:35:56	&	80	s	&	0.3953	&	$r$	&	16.57	$\pm$	0.04	\\
889: Jun 7	5:37:26	&	80	s	&	0.4203	&	$r$	&	16.75	$\pm$	0.08	\\
890: Jun 7	5:42:09	&	80	s	&	0.4989	&	$r$	&	16.69	$\pm$	0.07	\\
891: Jun 7	5:43:38	&	80	s	&	0.5236	&	$r$	&	16.60	$\pm$	0.06	\\
892: Jun 7	5:45:11	&	80	s	&	0.5494	&	$r$	&	16.64	$\pm$	0.08	\\
893: Jun 7	5:46:20	&	40	s	&	0.5686	&	$r$	&	16.78	$\pm$	0.12	\\
894: Jun 7	5:54:10	&	80	s	&	0.6992	&	$r$	&	17.25	$\pm$	0.08	\\
895: Jun 7	5:55:39	&	80	s	&	0.7239	&	$r$	&	17.37	$\pm$	0.08	\\
896: Jun 7	5:57:09	&	80	s	&	0.7489	&	$r$	&	17.37	$\pm$	0.10	\\
897: Jun 7	5:58:38	&	80	s	&	0.7736	&	$r$	&	17.54	$\pm$	0.12	\\
898: Jun 7	6:08:24	&	280	s	&	0.9365	&	$r$	&	17.65	$\pm$	0.11	\\
899: Jun 7	6:28:15	&	480	s	&	1.2672	&	$r$	&	17.69	$\pm$	0.06	\\
900: Jun 7	7:05:04	&	320	s	&	1.8810	&	$r$	&	18.44	$\pm$	0.08	\\
901: Jun 7	7:41:40	&	720	s	&	2.4909	&	$r$	&	18.77	$\pm$	0.09	\\
902: Jun 7	8:23:22	&	880	s	&	3.1859	&	$r$	&	19.04	$\pm$	0.15	\\
903: Jun 7	5:14:12	&	20	s	&	0.0331	&	$i$	&	14.75	$\pm$	0.05	\\
904: Jun 7	5:15:14	&	20	s	&	0.0503	&	$i$	&	14.05	$\pm$	0.04	\\
905: Jun 7	5:16:27	&	40	s	&	0.0706	&	$i$	&	14.20	$\pm$	0.04	\\
906: Jun 7	5:18:10	&	40	s	&	0.0992	&	$i$	&	14.76	$\pm$	0.05	\\
907: Jun 7	5:20:13	&	80	s	&	0.1333	&	$i$	&	15.01	$\pm$	0.04	\\
908: Jun 7	5:23:25	&	80	s	&	0.1867	&	$i$	&	15.61	$\pm$	0.05	\\
909: Jun 7	5:26:38	&	80	s	&	0.2403	&	$i$	&	15.99	$\pm$	0.07	\\
910: Jun 7	5:28:07	&	80	s	&	0.2650	&	$i$	&	16.23	$\pm$	0.08	\\
911: Jun 7	5:32:49	&	80	s	&	0.3433	&	$i$	&	16.34	$\pm$	0.10	\\
912: Jun 7	5:34:18	&	80	s	&	0.3681	&	$i$	&	16.43	$\pm$	0.08	\\
913: \enddata
914: \end{deluxetable}
915: 
916: \begin{deluxetable}{lcccc}
917: \tablecolumns{5}
918: \tablewidth{0pt}
919: \tablecaption{PROMPT Observations of the Afterglow of GRB 060607A (con't)}
920: \tablehead{\colhead{Mean Time (UT)} & \colhead{Exp. Time} & \colhead{Mean $\Delta$t (hr)} & \colhead{Filter}
921: & \colhead{Magnitude}}
922: \startdata
923: Jun 7	5:39:02	&	80	s	&	0.4469	&	$i$	&	16.48	$\pm$	0.08	\\
924: Jun 7	5:40:31	&	80	s	&	0.4717	&	$i$	&	16.53	$\pm$	0.07	\\
925: Jun 7	5:48:08	&	80	s	&	0.5986	&	$i$	&	16.68	$\pm$	0.09	\\
926: Jun 7	5:49:37	&	80	s	&	0.6233	&	$i$	&	16.60	$\pm$	0.11	\\
927: Jun 7	5:51:08	&	80	s	&	0.6486	&	$i$	&	16.79	$\pm$	0.10	\\
928: Jun 7	6:00:27	&	400	s	&	0.8038	&	$i$	&	17.33	$\pm$	0.07	\\
929: Jun 7	6:15:52	&	480	s	&	1.0609	&	$i$	&	17.46	$\pm$	0.07	\\
930: Jun 7	7:13:22	&	560	s	&	2.0190	&	$i$	&	18.32	$\pm$	0.09	\\
931: Jun 7	8:03:47	&	720	s	&	2.8594	&	$i$	&	18.94	$\pm$	0.19	\\
932: \enddata
933: \end{deluxetable}
934: 
935: \clearpage
936: 
937: \figcaption[f1.eps]{The $Bgri$ afterglow of GRB 060607A from 44 seconds 
938: to 3.2 hours taken with the PROMPT telescopes 
939: along with the $H$-band REM
940: light curve \citep{mvm+06}.  The solid lines present the 
941: best fit to the initial peak
942: at three minutes and two 
943: variations that occur at 42 and 64 minutes.
944: \label{0607_lc}}
945: 
946: \figcaption[f2.eps]{
947: The PROMPT $r$-band light curve plotted with the XRT x-ray afterglow
948: scaled to units of 10$^{4}$ counts.  Note that the time of the
949: variations in the optical correspond to a lapse in x-ray observations.
950: However, the x-ray light curve appears to flatten around this time and hence
951: does not undergo the same fluctuating
952: behavior as the optical light curve.
953: \label{0607_xrt}}
954: 
955: \figcaption[f3.eps]{The broadband spectral flux distribution
956: of the afterglow of 
957: GRB 060607A plotted at twelve minutes.  The solid lines are the fit to the
958: data; the dotted lines are the unextinguished model.  The high energy x-ray
959: flux is adopted from \citet{mvm+06}, and the plotted spectral index is 
960: taken from \citet{pgb06}.
961: \label{0607_spec}}
962: 
963: \clearpage
964: 
965: \setcounter{figure}{0}
966: 
967: \begin{figure}[tb]
968: \plotone{f1.eps}
969: \end{figure}
970: 
971: \begin{figure}[tb]
972: \plotone{f2.eps}
973: \end{figure}
974: 
975: \begin{figure}[tb]
976: \plotone{f3.eps}
977: \end{figure}
978: 
979: \clearpage
980: 
981: \bibliographystyle{apj}                 %citations using \citep and \citet
982: \bibliography{ms} 	                %Appropriate for use with BibTeX
983: 
984: \end{document}
985: