0712.2061/ms2.tex
1: \documentclass[preprint2]{aastex}
2: %\documentclass[12pt,preprint]{aastex}
3: 
4: %\usepackage[english]{babel}
5: %\usepackage {graphicx}
6: 
7: \def\ls{\vskip 12.045pt}
8: 
9: \newcommand{\cc}{\c{c}}
10: \newcommand{\cao}{\cc\~ao}
11: \newcommand{\coes}{\cc\~{o}es}
12: 
13: \begin {document}
14: 
15: \title{A two-armed pattern in flickering maps of the nova-like 
16:        variable UU~Aquarii\thanks{Based on observations made at the 
17:        Laborat\'orio Nacional de Astrof\'{\i}sica, CNPq, Brazil.}}
18: \shorttitle{flickering maps of UU Aqr}
19: 
20: \author{Raymundo Baptista} 
21: \affil{Departamento de F\'{i}sica , Universidade Federal de Santa 
22:      Catarina, Campus Trindade, 88040-900, Florian\'opolis, SC, Brazil}
23: \email{bap@astro.ufsc.br}
24: \and 
25: 
26: \author{Alexandre Bortoletto}
27: \affil{Instituto de Astronomia, Geof\'isica e Ci\^encias Atmosf\'ericas, 
28:      Universidade de S\~ao Paulo, Rua do Mat\~ao 1228, 05508-900, 
29:      S\~ao Paulo, SP, Brazil}
30: \email{abortoletto@astro.iag.usp.br}
31: 
32: \begin{abstract} 
33: We report the analysis of a uniform sample of 31 light curves of 
34: the nova-like variable UU~Aqr with eclipse mapping techniques.
35: The data were combined to derive eclipse maps of the average steady-light 
36: component, the long-term brightness changes, and low- and high-frequency 
37: flickering components.  The long-term variability responsible for the 
38: 'low' and 'high' brightness states is explained in terms of the response 
39: of a viscous disk to changes of 20-50 per cent in the mass transfer 
40: rate from the donor star.  
41: Low- and high-frequency flickering maps are dominated by emission from
42: two asymmetric arcs reminiscent of those seen in the outbursting dwarf 
43: nova IP~Peg, and are similarly interpreted as manifestation of a 
44: tidally-induced spiral shock wave in the outer regions of a large 
45: accretion disk.
46: The asymmetric arcs are also seen in the map of the steady-light
47: aside of the broad brightness distribution of a roughly steady-state 
48: disk. The arcs account for 25 per cent of the steady-light flux and
49: are a long-lasting feature in the accretion disk of UU~Aqr.
50: We infer an opening angle of $10\degr\pm3\degr$ for the spiral arcs. 
51: The results suggest that the flickering in UU~Aqr is caused by 
52: turbulence generated after the collision of disk gas with the 
53: density-enhanced spiral wave in the accretion disk.
54: \end{abstract}
55: 
56: \keywords {accretion, accretion disks -- binaries: eclipsing -- novae, 
57:  cataclysmic variables -- shock waves -- stars: activity -- stars: 
58: individual (UU Aquarii)}
59: 
60: 
61: \section{Introduction}
62: 
63: In Cataclysmic Variables (CVs) a late-type star overfills its Roche 
64: lobe and transfers matter to a companion white dwarf (WD) via an 
65: accretion column or disk. The CV zoo comprises low-mass transfer dwarf 
66: novae and high-mass transfer novae and nova-like systems. The light 
67: curve of these binaries show intrinsic brightness fluctuations 
68: (flickering) of 0.1-1 mag on timescales from seconds to tens of minutes, 
69: considered a basic signature of accretion (Warner 1995). 
70: When the Roche lobe of the WD is suficiently large there is room for 
71: the accretion disk to expand beyond the 3:1 resonance radius and the
72: tidal pull of the mass-donor star becomes relevant to the gas dynamics, 
73: giving rise to precessing elliptical rings (when the binary mass ratio 
74: $q=M_2/M_1 < q_\mathrm{crit}$, e.g. Whitehurst \& King 1991) and 
75: spiral shock waves (e.g., Steeghs 2001) in the outer disk regions.  
76: While theoretical expectations and numerical simulations suggest 
77: $q_\mathrm{crit}\simeq 0.3$ (e.g., Osaki 1996; Kunze, Speith \& Riffert
78: 1997; Murray, Warner \& Wickramasinghe 2000), an extensive study of
79: superhumps allowed Patterson et~al. (2005) to set an observational upper 
80: limit of $q_\mathrm{crit}=0.35 \pm 0.02$ for the creation of superhumps.
81: 
82: Earlier studies (Bruch 1992, 1996, 2000) led to the suggestion that 
83: there are mainly two sources of flickering in CVs, the stream-disk 
84: impact region at disk rim and a turbulent inner disk region in the 
85: vinicity of the WD (possibly the boundary layer), the relative 
86: importance of which varies from system to system. 
87: The spatially-resolved study of flickering in the dwarf nova 
88: V2051~Oph by Baptista \& Bortoletto (2004, hereafter BB04) revealed a 
89: more complex scenario, in which the low-frequency flickering is 
90: associated to an overflowing gas stream (possibly as a consequence of
91: unsteady mass transfer from the mass-donor star, e.g., Warner \& Nather 
92: 1971) and the high-frequency flickering is distributed over the surface
93: of the accretion disk -- possibly as a consequence of magnetohydrodynamic 
94: (MHD) turbulence or events of magnetic reconnection at the disk 
95: chromosphere (Geertsema \& Achterberg 1992; Kawagushi et~al. 2000). 
96: The identification of a disk component to the flickering and the
97: consequent estimation of the disk viscosity $\alpha$-parameter (Shakura
98: \& Sunyaev 1973) through the application of an MHD turbulence model, 
99: raised the expectation that the technique could be applied to measure
100: the accretion disk viscosity of other CVs. In particular, the tight 
101: correlation between the flickering disk component and the steady-light
102: emission in V2051~Oph (BB04) suggested that this could be the 
103: dominant source of flickering in nova-like systems, with their hot and 
104: bright accretion disks. An obvious next step would be to perform a 
105: spatially-resolved study of flickering on an eclipsing nova-like system
106: to test these ideas.
107: 
108: UU~Aqr is a bright eclipsing nova-like variable with an orbital period 
109: of 3.9~hr and a mass ratio $q=0.30\pm 0.07$ (Baptista, Steiner \& 
110: Cieslinski 1994; hereafter BSC). It shows conspicuous flickering 
111: activity (with flares which may reach 25\% of the total system light), 
112: long-term ($\sim 4\;yr$) changes of 0.3~mag in brightness (atributted 
113: to variations in the mass transfer rate from the mass-donor star, BSC), 
114: as well as $\sim 1$~mag changes on timescales of days (interpreted 
115: as being 'stunted' dwarf nova-type outbursts, Honeycutt, Robertson \& 
116: Turner 1998). It also displays superhumps in its light curve (Patterson 
117: et~al. 2005), suggesting the presence of a large, elliptical precessing 
118: accretion disk.
119: 
120: Here we report the analysis of a large sample of light curves of UU~Aqr
121: with eclipse mapping techniques to locate the sources of the flickering 
122: and to investigate the long-term brightness changes in this binary.
123: 
124: 
125: \section{Observations} \label{observa}
126: 
127: Time series of $B$-band CCD photometry of UU~Aqr were obtained with 
128: an EEV camera ($385\times 578$~pixels, $0.58\arcsec {\rm pixel}^{-1}$)
129: attached to the 0.6~m telescopes of the Laborat\'orio Nacional de
130: Astrof\'{\i}sica, in southern Brazil, from 1998 to 2001.  The CCD 
131: camera is operated in frame transfer mode, with negligible (13~ms) 
132: dead time between exposures. It has a Global Positioning System board 
133: that sets its internal clock to UTC to an accuracy better than 10~ms.  
134: The observations are summarized in Table~\ref{tab1}.
135: Columns 2 and 8 list the number of points in the light curve ($N_p$),
136: columns 3 and 9 give the exposure time in seconds ($\Delta t$). 
137: Columns 4 and 10 list the eclipse cycle number (E); the numbers in 
138: parenthesis indicate observations that, because of incomplete phase 
139: coverage or interruptions caused by clouds, do not cover the eclipse 
140: itself. Columns 6 and 12 give an estimate of the quality of the 
141: observations.  The seeing ranged from $1.5\arcsec$ to $2.5\arcsec$.
142: The observations comprise 31 light curves obtained with the same 
143: instrument and telescope, which ensures a high degree of uniformity 
144: to the data set. 
145: %
146: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% TABLE 1
147: \placetable{tab1}
148: 
149: Data reduction procedures included bias subtraction, flat-field 
150: correction, cosmic rays removal and aperture photometry extraction. 
151: Time series were constructed by computing the magnitude difference 
152: between the variable and a bright reference comparison star 
153: $77\arcsec$~E and $47\arcsec$~S of the variable with scripts based on 
154: the aperture photometry routines of the APPHOT/IRAF package
155:  \footnote{IRAF is distributed by National Optical Astronomy 
156:   Observatories, which is operated by the Association of Universities 
157:   for Research in Astronomy, Inc., under contract with the National 
158:   Science Foundation.}.  
159: Light curves of other comparison stars in the field were also computed 
160: in order to check the quality of the night and the internal consistency
161: and stability of the photometry over the time span of the observations.  
162: The magnitude and colors of the reference star were tied to the 
163: Johnsons-Cousins UBVRI systems (Bessell 1990) from observations of this 
164: star and of standard stars (Graham 1982; Stone \& Baldwin 1983) made 
165: on 4 photometric nights. We used the relations of Lamla (1981) to 
166: transform UBVRI magnitudes to flux units.  The B-band flux of the 
167: reference star was then used to transform the light curves of the 
168: variable from magnitude difference to absolute flux. We estimate that 
169: the absolute photometric accuracy of these observations is about 12 
170: per cent. On the other hand, the analysis of the light curves of field
171: comparison stars of brightness comparable to that of the variable 
172: indicates that the internal error of the photometry is less than 2 
173: per cent. The error in the photometry of the variable is derived from 
174: the photon count noise and is transformed to flux units using the 
175: same relation applied to the data. The individual light curves have 
176: typical signal-to-noise ratios of $S/N=$40-60 out-of-eclipse and 
177: $S/N=$10-20 at mid-eclipse.
178: 
179: 
180: \section{Data analysis} \label{analysis}
181: 
182: The light curves were phase-folded according to the linear ephemeris
183: (Borges 2005, private communication),
184: \begin{equation}
185: T_{mid}=HJD\; 2\,446\,347.2659 + 0.163\,804\,9430 \times E \; ,
186: \label{efem}
187: \end{equation}  
188: where $T_{mid}$ gives the WD mid-eclipse times.
189: Fig.\,\ref{fig1} shows the light curves of UU Aqr superimposed in 
190: phase. The upper frame depicts the light curves of a comparison star 
191: with the same brightness of UU~Aqr around mid-eclipse. The constancy 
192: of the comparison star flux over time indicates that all brightness 
193: variations seen in UU~Aqr are intrinsic to the variable. The scatter 
194: with respect to the mean level is significantly larger in UU Aqr and is 
195: caused by a combination of flickering and long term brightness changes. 
196: %
197: \placefigure{fig1}
198: 
199: We applied the `single' (Bruch 1996, 2000) and `ensemble' (Horne \& 
200: Stiening 1985; Bennie, Hilditch \& Horne 1996) methods to the set of
201: light curves of UU~Aqr to derive the orbital dependency of its 
202: steady-light, long-term brightness changes, low- and high-frequency
203: flickering components. The reader is referred to BB04 for a detailed
204: description and combined application of both methods.
205: 
206: UU~Aqr was in its 'high' brightness state during the 1998 and 2001
207: observations, and in its 'low' brightness state during the 1999 and 
208: 2000 runs. The curves of the 'high' and 'low' brightness states are 
209: identified in Fig.~\ref{fig1} by black and grey symbols, respectively. 
210: We remark that the average out-of-eclipse flux level increases steadily 
211: from the 1999 to the 2001 data and that the nominal separation between
212: the 'low' and 'high' brightness states is a rather arbitrary one. 
213: Because of the scatter produced by the strong flickering (with an 
214: average pear-to-peak amplitude of $\simeq 3$~mJy for both brightness
215: levels), there is an overlap in flux between the curves of the 'low'
216: and 'high' states.  In order to test the influence of the brightness 
217: state on the flickering behaviour, we applied the 'ensemble' method 
218: separately for the data of the 'high' and 'low' brightness states. 
219: We found no evidence of a dependency of the flux level or eclipse shape
220: of the flickering curve with brightness level in UU~Aqr. We therefore
221: combined all light curves for the following analysis. The difference 
222: in average brightness level seen in UU~Aqr along the observations is 
223: properly taken into account in the ensemble method by the curve of 
224: the long-term changes (see below, BB04).
225: 
226: In order to apply the 'ensemble' method, we define a reference 
227: out-of-eclipse flux $f_{ref}$ (the mean flux over the phase range
228: $0.15-0.80$) for each individual light curve and we divide the data
229: into a set of phase bins. A linear fit to the $f_i \times f_{ref}(i)$ 
230: diagram for each phase bin (e.g., Fig.~2 of BB04) yields an 
231: average flux (the steady-light component), an angular coefficient 
232: (which measures the long-term change) and a standard deviation with 
233: respect to the linear fit (the scatter curve, with added contributions 
234: from the Poisson noise and the flickering). 
235: We multiply the non-dimensional angular coefficients by the amplitude 
236: of the variation of the reference flux in the data set, 
237: $\Delta f_{ref}= 5.5\pm0.1$~mJy, to express the long-term changes in 
238: terms of the amplitude of the flux change per phase bin, 
239: $\Delta f_\nu(\phi)$. 
240: 
241: The average steady-light curve was subtracted from each individual 
242: light curve to remove the DC component and a Lomb-Scargle periodgram 
243: (Press et~al. 1992) was calculated. The periodgrams of all light curves 
244: were combined to yield a mean periodgram and a standard deviation with 
245: respect to the mean.  Fig.~\ref{fig2} shows the resulting average power 
246: density spectrum (PDS) binned to a resolution of 0.02 units in 
247: log(frequency). The PDS is well described by a power-law $P(f) \propto 
248: f^{-1.5}$. It becomes flat for $f_{low} < 0.15\;mHz \; (t_{low} > 111$ 
249: minutes) and disappears in the white noise for $f_{up}> 20\;mHz 
250: \;(t_{up}< 50\,s$). The slope of the PDS distribution is reminiscent
251: of those seen in other CVs, which can be well described by power laws 
252: with an average exponent $P(f) \propto f^{-2.0\pm0.8}$ (Bruch 1992).
253: %
254: \placefigure{fig2}
255: 
256: The 'ensemble' method samples flickering at all frequencies. But 
257: because of the power-law dependency of the flickering, an 'ensemble' 
258: curve is dominated by the low-frequency flickering components. 
259: On the other hand, the filtering process of the `single' method 
260: produces curves which sample the high-frequency flickering. The 
261: combination of both methods allows one to separate the low- (ensemble)
262: and high-frequency (single) components of the flickering.
263: 
264: The Achilles heel of the 'single' method is its difficulty in 
265: separating the high-frequency flickering from the rapid brightness
266: changes caused by the eclipse. 
267: % This problem imposes restrictions in the range of smoothing filter's 
268: % cutoff frequency applied in the method. 
269: In order to overcome this limitation, we subtracted the average 
270: steady-light curve from each individual light curve before applying the 
271: 'single' filtering process to eliminate the steep gradients produced by 
272: the eclipse in the light curve.
273: Our 'single' light curve includes flickering components with frequencies
274: $f_c > 2\;mHz$ (timescales shorter than $t_c=500\,s$, Fig.~\ref{fig2}). 
275: Single curves obtained with cut-off frequencies of $f_c= 3\;mHz\;
276: (t_c= 333\,s)$ and $f_c= 5\;mHz\;(t_c= 100\,s)$ show the same 
277: morphology of the lower cutoff frequency curve.  Because of the reduced
278: power, these other curves are noisier and will not be presented here.
279: 
280: The steady-light, long-term changes, and flickering curves were
281: analyzed with eclipse mapping techniques (Horne 1985; Baptista \& 
282: Steiner 1993) to solve for a map of the disk surface brightness 
283: distribution and for the flux of an additional uneclipsed component
284: in each case.  The uneclipsed component accounts for all light that is
285: not contained in the eclipse map (i.e., light from the secondary star
286: and/or a vertically extended disk wind). 
287: The reader is referred to Rutten, van Paradijs \& Tinbergen (1992) 
288: and Baptista, Steiner \& Horne (1996, hereafter BSH) for a detailed 
289: description of and tests with the uneclipsed component, and to 
290: Baptista (2001) for a recent review on the eclipse mapping method.
291: Out-of-eclipse brightness changes (not accounted for by the standard
292: eclipse mapping method) were removed from the light curves by fitting
293: a spline function to the phases outside eclipse, dividing the light 
294: curve by the fitted spline, and scaling the result to the spline 
295: function value at phase zero (e.g., BSH). 
296: 
297: Our eclipse map is a flat Cartesian grid of $51\times51$ pixels 
298: centered on the white dwarf with side $2 R_{L1}$ (where $R_{L1}$ is 
299: the distance from the disk center to the inner Lagrangian point L1).
300: The eclipse geometry is defined by the mass ratio $q$ and the 
301: inclination $i$, and the scale of the map is set by $R_{L1}$. We 
302: adopted $R_{L1}=0.744\;R_\odot$, $q=0.3$ and $i=78\degr$ (BSC), which
303: correspond to a white dwarf eclipse width of $\Delta\phi = 0.051$~cycle.
304: This combination of parameters ensures that the white dwarf is at the
305: center of the map.  The reconstructions were performed with a polar 
306: gaussian default function (Rutten et~al. 1992) with radial blur width 
307: $\Delta r= 0.02\;R_{L1}$ and azimuthal blur width $\Delta\theta= 
308: 30\degr$, and reached a final reduced chi-square $\chi_\nu^{2}\simeq 1$
309: for all light curves.  The uncertainties in the eclipse maps were 
310: derived from Monte Carlo simulations with the light curves using a 
311: bootstrap method (Efron 1982, Watson \& Dhillon 2001), generating a
312: set of 20 randomized eclipse maps (see Rutten et~al. 1992). These are
313: combined to produce a map of the standard deviations with respect to
314: the true map. A map of the statistical significance (or the inverse of
315: the relative error) is obtained by dividing the true eclipse map by 
316: the map of the standard deviations (Baptista et~al. 2005).
317: 
318: 
319: \section{Results}
320: 
321: Light curves and corresponding eclipse maps are shown in 
322: Fig.~\ref{fig3}.  For a better visualization of structures in the disk
323: brightness distributions, the asymmetric disk components are also 
324: shown. An asymmetric component is obtained by slicing the disk into a 
325: set of radial bins and fitting a smooth spline function to the mean 
326: of the lower half of the intensities in each bin. The spline-fitted 
327: intensity in each annular section is taken as the symmetric 
328: disk-emission component. The asymmetric disk component is then 
329: obtained by subtracting the symmetric disk from the original eclipse 
330: map (eg., Saito \& Baptista 2006). This procedure removes the baseline
331: of the radial distribution while preserving all azimuthal structure. 
332: %
333: \placefigure{fig3}
334: 
335: 
336: \subsection{Steady-light and long-term changes} \label{steady}
337: 
338: The steady-light light curve gives the flux per phase bin for the 
339: mid-reference flux level and represents the median steady brightness
340: level along the data set. Because it is obtained by combining 31 
341: light curves, it has high S/N and the corresponding eclipse map is
342: of high statistical significance (typically $>10\,\sigma$).
343: 
344: The eclipse map of the steady-light shows an extended brightness 
345: distribution peaking at disk center with two asymmetric arcs on
346: roughly opposite disk sides (Fig.~\ref{fig3}, top row).
347: The asymmetries are diluted by the dominant broad disk brightness 
348: distribution and only become clear in the asymmetric disk component. 
349: The arcs account for $\simeq 25$ per cent of the total flux of the 
350: steady light map. They are located at different radii, with the one 
351: in the trailing side (the lower disk hemisphere in the eclipse maps 
352: of Fig.~\ref{fig3}) being closer to disk center.  
353: The asymmetric arcs do not coincide with the WD at disk center or 
354: bright spot at disk rim. 
355: 
356: By transforming the intensities in the steady-light eclipse map into 
357: blackbody brightness temperatures (assuming a distance of 200\,pc to 
358: the binary, BSH) we find that the radial temperature distribution 
359: closely follows the $T\propto R^{-3/4}$ law for steady accretion in 
360: the outer disk and becomes flatter in the inner disk regions ($R 
361: \leq 0.2\,R_{L1}$), leading to estimated mass accretion rates of 
362: \.{M}$=10^{-9.0\pm0.2}\;M_\odot\;yr^{-1}$ at $R=0.1\,R_{L1}$ and 
363: \.{M}$=10^{-8.80\pm0.06}\;M_\odot\;yr^{-1}$ at $R=0.3\,R_{L1}$.
364: The brightness temperatures decrease steadily with radius, from 
365: $\simeq 13000\,K$ at $0.1\,R_{L1}$, to $\simeq 9400-7000\,K$ at
366: $(0.2-0.4)\,R_{L1}$ (the radial range at which the asymmetric arcs 
367: are located), and $\simeq 5500\,K$ at $0.5\,R_{L1}$.
368: We also find an uneclipsed component of $6.4 \pm 0.3$ per cent of 
369: the total steady-light flux.
370: The inferred brightness temperatures, uneclipsed component and mass
371: accretion rates are in good agreement with previous results (BSH, 
372: Baptista et~al. 2000; Vrielmann \& Baptista 2000). 
373: 
374: The curve of the long-term changes measures brightness changes on 
375: timescales longer than the orbital period. It allows us to visualize 
376: the differences in disk structure between the observed 'low' and 
377: 'high' brightness states of UU~Aqr.
378: 
379: The light curve of the long-term changes shows an eclipse with a 
380: pronouced shoulder at egress phases. The resulting eclipse map 
381: (Fig.~\ref{fig3}, second row from top) has a bright source at disk 
382: center and an azimuthally extended ($\Delta\theta\simeq 90\degr$) 
383: bright spot at disk rim, similar to those found in the eclipse maps 
384: of the 'high' state of BSH. The uneclipsed component is negligible.
385: This map tells us that the difference between the 'low' and 'high' 
386: brightness states is caused by an increase in the luminosity of the 
387: outer parts of the disk (as previously found by BSH) but also by a 
388: comparable increase in brightness of the innermost disk regions.
389: 
390: For a fixed distance, the intensities in the eclipse map scale 
391: linearly with the flux in the light curve (see, e.g., Baptista 2001).
392: Because we choose to express the long-term changes curve in terms of 
393: the amplitude of the flux variation (Sect.~\ref{analysis}), the 
394: intensities in the corresponding eclipse map are given in terms of the
395: amplitude of the variation in intensity between the minimum and maximum
396: brightness states in the data set, $\Delta I_j$ (where $j$ refers to
397: each pixel in the eclipse map). Thus, eclipse maps representing the
398: minimum and maximum brightness distributions can be obtained by 
399: adding/subtracting the appropriate proportion of the long-term 
400: changes map to/from the steady-light map, $\bar{I_j}$,
401: %
402: \begin{mathletters}
403: \begin{equation}
404: I_j(\mathrm{max})= \bar{I_j} + \frac{1}{2} \Delta I_j \,\,\, , 
405: \end{equation}
406: \begin{equation}
407: I_j(\mathrm{min})= \bar{I_j} - \frac{1}{2} \Delta I_j \,\,\, .
408: \end{equation}
409: \end{mathletters}
410: %
411: As expected, the resulting minimum and maximum brightness maps 
412: (Fig.~\ref{fig4}) are similar to the $B$-band eclipse maps of the 
413: 'low' and 'high' brightness states of BSH (see their Fig.\,3).
414: %
415: \placefigure{fig4}
416: 
417: We now turn our attention to the interpretation of the structures 
418: seen in the long-term changes map.
419: BSH suggested that the azimuthally extended spot seen in the 'high'
420: brightness state reflects long-term changes in luminosity caused by 
421: variations in mass input rate at the outer disk. 
422: Baptista et~al. (2000) noted that it could alternatively be the
423: signature of an elliptical (precessing) disk similar to those 
424: possibly present in SU~UMa stars during super-outbursts.
425: 
426: In order to test the second possibility, we searched for the presence
427: of superhumps in our data set. We combined the light curves for a 
428: given year of observations and computed Lomb-Scargle periodograms
429: after removing the eclipses from the data. We find no statistically
430: significant periodicities in the combined light curves at the range
431: of frequencies near the orbital frequency. The only exception is
432: the data of 1999 (JD 2,451,403-2,451,404, when UU~Aqr was on its 
433: lowest brightness level along our data set), which presents a clear 
434: signal at the orbital frequency. A sine fit to the data shows that
435: the peak occurs at phases $\phi_{max}=0.87-0.92$~cycle on all orbits, 
436: indicating that we are seeing the orbital hump and not a superhump.
437: We remark that, because our observations focus on the eclipse and 
438: frame relatively short time intervals (of at most 2-3 consecutive 
439: orbits) on each night, they are not as sensitive to the presence of 
440: superhumps as the much more intensive and well sampled data of the
441: campaign by Patterson et~al. (2005).
442: Therefore, the lack of positive identification of superhumps in
443: our data indicates that, if present, the superhump signal was
444: of low amplitude/intensity and would hardly be able to account for 
445: the extended bright spot seen in the long-term changes map.
446: We are thus left with the explanation of BSH.
447: 
448: Can the bright source at inner disk also be explained in terms of
449: changes in mass input rate?  In order to answer this question, we took 
450: a reference steady-state disk model with mass accretion rate 
451: $\hbox{\.{M}}_\mathrm{ref}$ and computed the difference between the 
452: (blackbody) brightness distributions of a test steady-state model with 
453: $\hbox{\.{M}}_\mathrm{test}= \beta\, \hbox{\.{M}}_\mathrm{ref}$ 
454: ($\beta>1$) and the reference model. For an opaque steady-state disk, 
455: the change in effective temperature, $\partial T$, caused by a change 
456: in mass accretion rate, $\partial$\.{M}, has the same radial dependency
457: of the effective temperature (see, e.g., Frank, King \& Raine 1992),
458: %
459: \begin{equation}
460: \frac{\partial T}{\partial\hbox{\.{M}}} \propto R^{-3/4} 
461: \left[ 1 - \left( \frac{R_{wd}}{R} \right)^{1/2} \right]^{1/4} \,\, ,
462: \end{equation}
463: %
464: where $R_{wd}$ is the WD radius.
465: Thus, the difference in temperature (and the corresponding difference in 
466: blackbody intensity) between two steady-state disk models increases with 
467: decreasing radius and peaks near disk center (at $R= 49/36\,R_{wd}$). 
468: The difference in intensity also scales with ${\partial\hbox{\.{M}}}$.
469: In searching for the pair of ($\hbox{\.{M}}_\mathrm{ref}, \beta$) values 
470: of best fit to the observed brightness distribution we find 
471: $\hbox{\.{M}}_\mathrm{ref}= 10^{-9.1\pm 0.1}\;M_\odot\;yr^{-1}$ and 
472: $\beta= 1.35\pm 0.15$.
473: Fig.~\ref{fig5} shows that the radial distribution of the central source
474: of the long-term changes map is consistent, within the uncertainties, 
475: with the difference in intensity expected for an increase by 20-50 per 
476: cent in mass accretion rate of a steady-state disk with $10^{-9.1}\;
477: M_\odot\;yr^{-1}$. Because the radial temperature distribution of the
478: steady-light map is actually flatter than the $T \propto R^{-3/4}$ law
479: of steady-state disks, these results should be considered illustrative.
480: Nevertheless, the inferred range of mass accretion rates ($10^{-9.1} -
481: 10^{-8.9}\,M_\odot\;yr^{-1}$) is in line with the values for the 'low'
482: and 'high' brightness states found by BSH.
483: %
484: \placefigure{fig5}
485: 
486: Because in a steady-state disk the mass accretion rate reflects the
487: mass transfer rate, $\hbox{\.{M}}_2$, we may interpret the map of the 
488: long-term changes in terms of the response of a high-viscosity disk to 
489: changes in $\hbox{\.{M}}_2$ of about 20-50 per cent.  
490: When $\hbox{\.{M}}_2$ increases, the luminosity of the bright spot at
491: disk rim increases, as well as that of the inner disk regions -- as a
492: consequence of the increase of mass accretion through a disk close to
493: a steady-state.
494: 
495: 
496: \subsection{low- and high-frequency flickering} \label{flick}
497: 
498: The 'ensemble' and 'single' curves show a double-stepped eclipse 
499: reminiscent of the occultation of the two-armed spiral structure seen 
500: in eclipse maps of the dwarf nova IP~Peg in outburst (e.g., Baptista,
501: Haswell \& Thomas 2002; Baptista et~al. 2005), and lead to similar 
502: two-armed asymmetric brightness distributions (Fig.~\ref{fig3}, the two 
503: lowermost rows). The solid contour line overploted on each eclipse 
504: map depicts the 3-$\sigma$ confidence level region as derived from the 
505: map of statistical significance in each case (Sect.~\ref{analysis}).
506: The asymmetric arcs are at or above the 3-$\sigma$ confidence level in 
507: both maps. Vertical tick marks in the 'ensemble' panel of Fig.~\ref{fig3} 
508: indicate the eclipse ingress/egress phases of the two bright arcs 
509: (labeled '1' and '2').  Their location is depicted in the asymmetric 
510: component of the ensemble map. A simple comparison reveals that these are 
511: the same asymmetric arcs seen in the steady-light map. The major 
512: difference is that the arcs dominate the emission in the flickering maps 
513: (the asymmetric component account for 53 and 41 per cent of the total 
514: flux, respectively for the 'ensemble' and 'single' maps).
515: As already noted, they do not coincide with the WD at disk center nor 
516: with the bright spot at disk rim.
517: 
518: Although it is possible to center the eclipse of source 1 by applying 
519: a phase shift to all light curves, this would lead to physically 
520: unplausible brightness distributions for the steady-light and long-term
521: changes maps -- with highly asymmetric brightness distributions where 
522: the main sources fall at positions which cannot be associated with 
523: either the WD, the bright spot or gas stream (e.g., the azimuthally
524: extended spot in the long-term changes map would fall at the edge of the
525: primary Roche lobe, far away from the gas stream trajectory). 
526: It is also not possible to interpret the observed asymmetries in terms 
527: of enhanced emission from an elliptical outer disk ring because the 
528: asymmetries lie well inside the disk, far from its edge.
529: Given the similarities with the IP~Peg eclipse maps (e.g., see Fig.~3
530: of Baptista et~al. 2002) and the lack of plausible alternative 
531: explanations, we interpret the asymmetries in the flickering and 
532: steady-light maps as consequence of tidally-induced spiral shock waves 
533: in the accretion disk of UU~Aqr (e.g., Sawada et~al. 1986).
534: 
535: Arc 1 is in the trailing side of the disk at $R_{s1}= 0.20\pm 0.05\;
536: R_{L1}$; Arc 2 is in the leading side of the disk and is farther away 
537: from disk center, $R_{s2}= 0.32 \pm 0.05\;R_{L1}$.
538: The two arcs have azimuthal extent $\Delta\theta \simeq 110 \degr$ 
539: and radial extent $\Delta R \simeq 0.2\,R_{L1}$.
540: Baptista et~al. (2005) devised a way to estimate the opening angle of 
541: the spirals from the azimuthal intensity distribution of the eclipse
542: maps. We applied the same method to the flickering maps to estimate an 
543: opening angle of $\phi=10\degr \pm 3\degr$, indicating that the spiral 
544: arms in UU~Aqr are more tightly wound than in IP~Peg at outburst 
545: ($\phi= 14\degr - 34\degr$, Baptista et~al. 2005). They are also
546: systematically closer to disk center than the arms seen in IP~Peg (at 
547: average distances of 0.30 and $0.55\,R_{L1}$, see Baptista et~al. 2005).
548: Because the opening angle of the spiral arms scales with the disk 
549: temperature (e.g., Steeghs \& Stehle 1999), this suggests that the 
550: outer accretion disk of UU~Aqr is cooler than that of IP~Peg in 
551: outburst.
552: 
553: We find uneclipsed components of $13\pm3$ and $17\pm7$ per cent of 
554: the total flux, respectively for the 'ensemble' and 'single' maps. 
555: This suggests that a sizeable part of the flickering may arise from 
556: outside the orbital plane, perhaps in a vertically-extended disk 
557: chromosphere + wind.
558: 
559: The 'ensemble' map samples flickering at all frequencies while the
560: 'single' map contains the high-frequency (timescales $<500\,s$) 
561: flickering components. It is possible to separate the contribution of
562: the low-frequency flickering by subtracting the 'single' map from the 
563: 'ensemble' map.  
564: 
565: Fig.~\ref{fig6} compares the relative amplitude of the low- ('ensemble' 
566: - 'single') and high-frequency ('single') flickering components in UU~Aqr. 
567: The radial run of the relative amplitudes are obtained by dividing the 
568: average radial intensity distribution of these two flickering maps by 
569: that of the steady-light. Dashed and dotted lines show the 1-$\sigma$ 
570: limits on the average amplitude, respectively for the low- and 
571: high-frequency flickering. The large uncertainties reflect the scatter 
572: introduced by the asymmetric arcs in the radial bins.
573: The two distributions are comparable, within the uncertainties.
574: Flickering is negligible in the inner disk. The amplitude of the 
575: low-frequency flickering increases monotonically with radius, reaching 
576: 6 per cent of the total light at $0.3\,R_{L1}$. The amplitude of the 
577: high-frequency flickering also increases with radius and peaks at the 
578: location of the spiral arms ($\simeq 0.3\,R_{L1}$). 
579: The distributions are not reliable for $R\geq 0.45\,R_{L1}$ because
580: of the reduced statistical significance of the flickering maps and the
581: rapidly declining intensities in the steady-light map.
582: The lower panel of Fig.~\ref{fig6} depicts the ratio of the 'single'
583: to the 'ensemble' distributions and measures the contribution of the
584: high-frequencies ('single') to the total flickering.
585: The high-frequency flickering accounts for a roughly constant fraction
586: of $\simeq 50$ per cent of the flickering signal at all radii. 
587: There is no apparent difference in radial behaviour between the low- 
588: and high-frequency flickering components, indicating that they not only 
589: arise from the same location but are also produced by the same physical
590: mechanism.
591: 
592: We further notice that our data cover a time interval of about 4~yr
593: and that the eclipse maps yield the average behaviour over this time
594: scale -- with the implication that the observed spiral shocks are a 
595: long-lasting feature in the accretion disk of UU~Aqr.
596: 
597: 
598: \section{Discussion}
599: 
600: BB04 found two independent sources of flickering in their study of
601: the dwarf nova V2051~Oph: the low-frequency flickering is associated
602: to inhomogeneities in the mass transfer process, while the high-frequency
603: flickering originates in the accretion disk, possibly as a consequence
604: of MHD turbulence (Geertsema \& Achterberg 1992) or events of magnetic 
605: reconnection in the disk surface (Kawagushi et~al. 2000).
606: 
607: Contrary to the suggestion of Bruch (2000), UU~Aqr shows no evidence of 
608: flickering arising from a turbulent inner disk or from the bright spot. 
609: And, in contrast to V2051~Oph, it shows no disk-related flickering 
610: component. However, its high- and low-frequency flickering have the same 
611: origin. They are produced in a two-armed pattern reminiscent of the 
612: tidally-induced spiral shocks seen in outbursting accretion disks of 
613: dwarf novae (Steeghs, Harlaftis \& Horne 1997; Baptista et~al. 2002).
614: These shocks are produced by tidal effects when the disk expands beyond 
615: the 3:1 resonance radius.  In a dwarf novae these spiral shocks are only
616: seen during outburst, when the disk becomes hot and large.  In a 
617: nova-like system, the spiral shocks may be a permanent feature if the
618: hot disk is large enough for tidal effects to become relevant.
619: The presence of spiral structures in the steady-light and flickering maps
620: of UU~Aqr is a likely indication that its accretion disk is large enough
621: for the tidal pull of the mass-donor star to be relevant for the gas
622: dynamics in the outer disk regions. This is in line with the detection 
623: of long-lasting superhumps in this binary (Patterson et~al. 2005) -- 
624: the tidal influence that leads to spiral density waves may also induce
625: elliptical orbits in the outer disk regions. Because the spiral arms 
626: account for a small fraction of the steady-light and are not related to
627: the broad and brighter steady emission centered in the disk, one might 
628: conclude that they are not the dominant source of viscous dissipation 
629: and angular momentum removal in the UU~Aqr accretion disk.
630: 
631: Why do the asymmetric arcs flicker? We discuss three possibilities.
632: One may consider that flickering is a consequence of unsteady 
633: dissipation of energy from a clumpy gas stream as it hits the two-armed
634: spiral density wave in the disk (i.e., a mass transfer origin for the 
635: flickering, as proposed by Warner \& Nather 1971).
636: In this case the clumpy gas stream should also lead to detectable 
637: flickering at the location of the bright spot, where it first hits the
638: accretion disk rim before reaching the spiral arms.  However, while the
639: bright spot is a significant light source in the long-term changes map
640: (indicating that the stream-disk impact occurs at $R\simeq 0.6\,R_{L1}$,
641: farther out in the disk than the observed radial position of the spiral
642: arms) it gives no contribution to the flickering. We may conclude that
643: there is no evidence of clumpiness in the infalling gas stream and,
644: therefore, we shall exclude this as a viable explanation. 
645: A second possibility is to consider that flickering arises from
646: reprocessing at tidally-induced and vertically-thickened disk regions
647: of unsteady irradiation from the boundary layer (a boundary layer/WD
648: flickering, Bruch 1992). The problem with this explanation is similar to 
649: that of the previous one: it would be hard to explain why we do not see 
650: optical flickering directly from the innermost disk regions.
651: The third and most promising possibility is to assume that the observed 
652: flickering is the consequence of turbulence generated by the shock of 
653: disk gas as it passes through the tidally-induced spiral density waves 
654: (i.e., a local origin for the flickering).
655: Further numerical simulations of spiral shocks in accretion disks would
656: be useful to verify the turbulent nature of the after-shock gas and to
657: test if such turbulence could generate the observed power-law dependency
658: [$P(f) \propto f^{-1.5}$] of the resulting energy dissipation fluctuations.
659: 
660: Assuming that the disk gas moves in Keplerian orbits around the $0.67\;
661: M_\odot$ WD (BSC), the locations of the asymmetric arcs corresponds to 
662: Doppler velocities of $900-1050 \;km\,s^{-1}$ and $700-800\;km\,s^{-1}$, 
663: respectively for arcs 1 and 2. Given that the after-shock gas is expected
664: to have sub-Keplerian velocities ($\simeq 15-40$ per cent lower, see 
665: Steeghs \& Stehle 1999; Baptista et~al. 2005), these structures may appear 
666: as arcs of enhanced emission at velocities $\simeq 450-650\;km\,s^{-1}$ 
667: in the upper left (arc 2) and lower right (arc 1) quadrant on a Doppler 
668: tomogram. Kaitchuck et~al. (1998) and Hoard et~al. (1998) report that
669: UU~Aqr disk line-emission is largely asymmetric. Their tomograms show 
670: regions of enhanced emission which may be interpreted as arising from a
671: two-armed spiral pattern in the disk. The asymmetry corresponding to arc 
672: 1 is clearly seen (e.g., see Fig.\,6 of Hoard et~al. 1998 and Fig.\,15 of
673: Kaitchuck et~al. 1998), while that related to arc 2 may be blended with 
674: and hidden by the emission from the gas stream and bright spot impact site. 
675: This latter effect may help to explain the large mass ratio ($q=0.86$) 
676: inferred by Kaitchuck et~al. (1998) by fitting the gas stream trajectory
677: to the asymmetry in the upper left quadrant of their Doppler tomograms.
678: 
679: 
680: \section{Summary}
681: 
682: Our investigation of the sources of variability in UU~Aqr indicates that
683: the long-term changes giving rise to the 'low' and 'high' brightness 
684: states can be accounted for by changes in mass transfer rate of 20-50
685: per cent. A high S/N steady-light light curve reveal the presence of
686: long-lasting (at least over a 4 yr period of the observations) 
687: asymmetric arcs in the accretion disk aside of the broad brightness
688: distributions of a roughly steady-state disk. The arcs are interpreted 
689: as the consequence of tidally-induced spiral shocks in an extended and 
690: hot accretion disk. The spiral arms account for 25 per cent of the 
691: steady-light flux and are the dominant source of flickering, both at 
692: low- and high-frequencies. They are more tightly wound than the spiral 
693: shocks found in the outbursting dwarf nova IP~Peg. The observed 
694: flickering shows a power spectrum density with an $f^{-1.5}$ 
695: power-law dependency, and is best explained as resulting from turbulence 
696: generated after the collision of disk gas with the density-enhanced 
697: spiral wave in the accretion disk.
698: There is no evidence of flickering originating in the bright spot at
699: disk rim or in the innermost disk regions around the WD.
700: 
701: 
702: \acknowledgements
703: 
704: This work was partially supported by CNPq/Brazil through research grant
705: 62.0053/01-1-PADCT III/Milenio.
706: RB acknowledge financial support from CNPq/Brazil through grants no. 
707: 300\,354/96-7 and 200\,942/2005-0.
708: 
709: 
710: \begin{thebibliography}{}
711: 
712: \bibitem{bap01}
713: Baptista, R. 2001. in Astrotomography, Indirect Imaging Methods in 
714: Observational Astronomy, Lecture Notes in Physics 573, ed.\ H. M. J. 
715: Boffin, D. Steeghs, \& J. Cuypers (Berlin: Springer), 307
716: 
717: \bibitem{bb04} 
718: Baptista, R., \& Bortoletto, A., 2004. \aj, 411, 128 (BB04)
719: 
720: \bibitem{bht02}
721: Baptista, R., Haswell, C. A., \& Thomas, G., 2002. \mnras, 334, 198
722: 
723: \bibitem{bmhms}
724: Baptista, R., Morales-Rueda, L., Harlaftis, E. T., Marsh, T. R., \&
725: Steeghs, D., 2005. \aap, 444, 201
726: 
727: \bibitem{bssh00}
728: Baptista, R., Silveira, C., Steiner, J. E., \& Horne, K., 2000. 
729: \mnras, 314, 713
730: 
731: \bibitem{bs93} 
732: Baptista, R., \& Steiner, J. E. 1993. \aap, 277, 331
733: 
734: \bibitem{bsc94}
735: Baptista, R., Steiner, J. E., \& Cieslinski, D., 1994. \apj, 433, 
736: 332 (BSC)
737: 
738: \bibitem{bsh96}
739: Baptista, R., Steiner, J. E., \& Horne, K., 1996. \mnras, 282, 99 (BSH)
740: 
741: \bibitem{bennie}
742: Bennie, P. J., Hilditch, R., \& Horne, K., 1996. in IAU Colloq.\ 
743: 158, Cataclysmic Variables and Related Objects, ed.\ A. Evans \& 
744: J. H. Wood (Dordrecht: Kluwer), 33
745: 
746: \bibitem {bessell} 
747: Bessell, M. A. 1990. \pasp, 102, 1181
748: 
749: \bibitem{bruch92}
750: Bruch, A., 1992. \aap, 266, 237
751: 
752: \bibitem{bruch96}
753: Bruch, A., 1996. \aap, 312, 97
754: 
755: \bibitem{bruch00}
756: Bruch, A., 2000. \aap, 359, 998
757: 
758: \bibitem{boots} Efron, B., 1982. The Jackknife, The Bootstrap and 
759: Other Resampling Plans (SIAM, Philadelphia 1982)
760: 
761: \bibitem{acpower}
762: Frank, J., King, A., \& Raine D. 1992. Accretion Power in Astrophysics
763: (2nd.\ edition, Cambridge: Cambridge Univ.\ Press)
764: 
765: \bibitem{graham}
766: Graham, J. A., 1982. \pasp, 94, 244
767: 
768: \bibitem {ga}
769: Geertsema, G. T. \& Achterberg, A. 1992. \aap, 255, 427
770: 
771: \bibitem {hoard}
772: Hoard, D. W., Still, M. D., Szkody, P., Smith R. C., Buckley, D. A. H.,
773: 1998. \mnras, 294, 689
774: 
775: \bibitem{horne85} 
776: Horne, K. 1985. \mnras, 213, 129
777: 
778: \bibitem{hs85}
779: Horne, K., \& Stiening, R. F., 1985. \mnras, 216, 933
780: 
781: \bibitem{honey98} 
782: Honeycutt, R. K., Robertson, J. W., \& Turner, G. W., 1998. \aj, 115, 
783: 2527
784: 
785: \bibitem{kait}
786: Kaitchuck, R. H., Schlegel, E. M., White II, J. C., Mansperger, C. S.,
787: 1998. \apj, 499, 444
788: 
789: \bibitem {kawa} 
790: Kawaguchi, T., Mineshige, S., Machida, M., Matsumoto, R. \& Shibata K. 
791: 2000. PASJ, 52, L1
792: 
793: \bibitem {kunze}
794: Kunze, S., Speith, R., \& Riffert, H., 1997. \mnras, 289, 889
795: 
796: \bibitem {murray}
797: Murray, J. R., Warner, B., \& Wickramasinghe, D. T., 2000. \mnras, 
798: 315, 707
799: 
800: \bibitem {lamla} 
801: Lamla, E. 1981. in Landolt-B\"{o}rnstein - Numerical Data and Functional 
802: Relationships in Science and Technology, Vol.\,2, eds.\ K. Schaifers 
803: \& H. H. Voigt (Berlin: Springer-Verlag)
804: 
805: \bibitem {osaki}
806: Osaki, Y., 1996. \pasp, 108, 39
807: 
808: \bibitem{patterson}
809: Patterson, J., et~al., 2005. \pasp, 117, 1204
810: 
811: \bibitem {numrec} 
812: Press, W. H., Flannery, B. P., Teukolsky, S. A., \& Vetterling, W. T., 
813: 1992. Numerical Recipes in C (Cambridge: Cambridge University Press)
814: 
815: \bibitem{rutten} 
816: Rutten, R. G. M., van Paradijs, J. \& Tinbergen, J. 1992. \aap, 260, 213
817: 
818: \bibitem{saito} 
819: Saito, R. K, \& Baptista, R. 2006. \aj, 131, 2185
820: 
821: \bibitem{sawada}
822: Sawada, K., Matsuda, T., \& Hachisu, I., 1986. \mnras, 219, 75
823: 
824: \bibitem {ss} 
825: Shakura, N. I. \& Sunyaev, R. A. 1973. \aap, 24, 337
826: 
827: \bibitem{spirals}
828: Steeghs, D. 2001. in Astrotomography, Indirect Imaging Methods in 
829: Observational Astronomy, Lecture Notes in Physics 573, ed.\ H. M. J. 
830: Boffin, D. Steeghs, \& J. Cuypers (Berlin: Springer), 45
831: 
832: \bibitem{steeghs}
833: Steeghs, D., Harlaftis, E. T., \& Horne, K, 1997. \mnras, 290, L28
834: 
835: \bibitem {stst}
836: Steeghs, D., \& Stehle, R., 1999, \mnras, 307, 99
837: 
838: \bibitem{sb} 
839: Stone, R. P. S., \& Baldwin, J. A., 1983. \mnras, 204, 347
840: 
841: \bibitem{w95} 
842: Warner, B. 1995. Cataclysmic Variable Stars (Cambridge: Cambridge 
843: University Press)
844: 
845: \bibitem {wn} 
846: Warner, B. \& Nather, R. E. 1971. \mnras, 152, 219
847: 
848: \bibitem{roche} 
849: Watson, C. A., \& Dhillon, V. S., 2001. \mnras, 326, 67
850: 
851: \bibitem{precess}
852: Whitehurst, R., \& King, A., 1991. \mnras, 249, 25
853: 
854: \bibitem{vb00}
855: Vrielmann, S., \& Baptista, R., 2002. Astron. Nach., 323, 75
856: 
857: \end{thebibliography}
858: 
859: \clearpage
860: 
861: %%%%%%%%%% TABLES
862: %
863: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% TABLE 1
864: % \tableline\tableline
865: %  HJD start & $N_p$ & $\Delta\,t$ & E & phase & Quality &
866: %  HJD start & $N_p$ & $\Delta\,t$ & E & phase & Quality \\ [-0.5ex]
867: % (2.450.000+) && (s) & (cycle) & range && 
868: % (2.450.000+) && (s) & (cycle) & range & \\ 
869: % \tableline
870: \begin{deluxetable}{crcccccrcccc}
871: \tabletypesize{\scriptsize}
872: \tablecaption{Journal of the observations\label{tab1}}
873: \tablewidth{0pt}
874: \tablehead{
875: \colhead{HJD start} & \colhead{$N_p$} & \colhead{$\Delta\,t$} & 
876: \colhead{E \tablenotemark{a}} & \colhead{phase} & \colhead{Quality 
877: \tablenotemark{b}} & 
878: \colhead{HJD start} & \colhead{$N_p$} & \colhead{$\Delta\,t$} & 
879: \colhead{E \tablenotemark{a}} & \colhead{phase} & \colhead{Quality 
880: \tablenotemark{b}} \\
881: \colhead{(2.450.000+)} && \colhead{(s)} & \colhead{(cycle)} & 
882: \colhead{range} &&
883: \colhead{(2.450.000+)} && \colhead{(s)} & \colhead{(cycle)} & 
884: \colhead{range}
885: }
886: \startdata
887:  1047.6259 &  368 & 10 & (28734) & $+0.24,+0.50$ & A & 
888:  1782.5171 &  999 & 10 & 33227 & $-0.22,+0.48$ & B \\
889:  1047.6683 & 1446 & 10 & 28735 & $-0.50,+0.50$ & A &  
890:  1782.6490 &  433 & 10 & (33228) & $-0.42,-0.10$ & C \\ 
891:  1050.4957 &  506 & 10 & 28752 & $-0.22,+0.15$ & A &  
892:  1787.6155 &  347 & 10 & 33258 & $-0.05,+0.25$ & C \\ 
893:  1050.6815 &  339 & 10 & 28753 & $-0.08,+0.17$ & A &  
894:  2136.5369 &  649 & 10 & 35391 & $-0.03,+0.45$ & B/C \\ 
895:  1052.4647 & 1179 &  7 & 28764 & $-0.18,+0.41$ & B &
896:  2136.6232 &  538 & 10 & (35392) & $-0.50,+0.50$ & B/C \\ 
897:  1052.5936 & 1322 &  7 & 28765 & $-0.39,+0.43$ & B &  
898:  2137.4796 & 1079 & 10 & 35397 & $-0.27,+0.50$ & B \\ 
899:  1053.5675 & 1397 &  8 & 28771 & $-0.44,+0.35$ & B/C &  
900:  2137.6048 & 1413 & 10 & 35398 & $-0.50,+0.50$ & B \\ 
901:  1053.7460 &  201 & 20 & (28772) & $-0.35,-0.06$ & B/C &  
902:  2138.4690 &  676 & 15 & 35403 & $-0.22,+0.50$ & B \\ 
903:  1054.6388 &  169 & 15 & (28777) & $+0.11,+0.29$ & C &  
904:  2138.5863 &  756 & 15 & 35404 & $-0.50,+0.50$ & B \\ 
905:  1055.5546 &  265 & 10 & (28783) & $-0.29,-0.06$ & C &  
906:  2138.7499 &  550 & 15 & 35405 & $-0.50,+0.09$ & B \\ 
907:  1403.4937 &  898 & 10 & 30910 & $-0.27,+0.37$ & B &  
908:  2139.4636 &  590 & 15 & 35409 & $-0.14,+0.50$ & B \\
909:  1403.6550 &  885 & 10 & 30911 & $-0.28,+0.39$ & B & 
910:  2139.5678 &  886 & 15 & 35410 & $-0.50,+0.50$ & A \\
911:  1404.4772 & 1068 & 10 & 30916 & $-0.26,+0.50$ & B & 
912:  2139.7313 &  604 & 15 & 35411 & $-0.50,+0.14$ & A \\
913:  1404.6009 & 1242 & 10 & 30917 & $-0.50,+0.50$ & B &
914:  2140.6193 &  380 & 15 & 35416 & $-0.07,+0.33$ & C \\ 
915:  1404.7644 &  563 & 10 & (30918) & $-0.50,-0.10$ & B &
916:  2141.5954 &  192 & 15 & 35422 & $-0.10,+0.10$ & B/C \\ 
917: &&&&&& 2141.7556 & 346 & 15 & 35423 & $-0.13,+0.12$ & B/C \\
918: [0.5ex]
919: \enddata
920: \tablenotetext{a}{ with respect to the ephemeris of eq.~(\ref{efem}).}
921: \tablenotetext{b}{ sky conditions: A= photometric (main comparison 
922: stable), B= good (some sky variations), C= poor (large variations 
923: and/or clouds).} 
924: \end{deluxetable}
925: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
926: 
927: \clearpage
928: 
929: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% FIG.1
930: \begin{figure}
931: %\center
932: %\includegraphics[scale=0.6,angle=270]{f1.eps}
933: \caption{Light curves of UU~Aqr (lower panel) and of a comparison star 
934:   (upper panel) superimposed in phase. The data of the 'high' and 'low'
935:   brightness states are plotted with black and grey symbols, 
936:   respectively. (SEE ATTACHED JPG FILE) }
937: \label{fig1}
938: \end{figure}
939: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
940: 
941: %\clearpage
942: 
943: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% FIG.2
944: \begin{figure}
945: \includegraphics[angle=270,scale=0.6]{f2.eps}
946: \caption{Average power density spectrum. Dotted lines show the $1\sigma$ 
947:   limits on the average power. The best-fit power-law $P(f) \propto 
948:   f^{-1.5}$ is shown as a dashed line. A vertical tick marks the
949:   low-frequency cutoff of the filtering process applied to derive the
950:   'single' scatter curve. }
951: \label{fig2}
952: \end{figure}
953: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
954: 
955: \clearpage
956: 
957: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% FIG.3
958: \begin{figure*}
959: \center
960: \includegraphics[bb= 1cm 0.5cm 19.5cm 24.5cm,angle=270,scale=0.7]{f3.eps}
961: \caption{Left-hand panels: data (dots) and model (solid lines) light 
962:  curves of UU Aqr. From top to bottom: the steady-light component, the 
963:  curve of the long-term changes (secular), the 'ensemble' and the 
964:  'single' flickering curves.  Vertical dotted lines mark the 
965:  ingress/egress and mid-eclipse phases of the white dwarf. A dashed 
966:  curve depicts the error bars in each case. A dotted curve in the 
967:  'ensemble' panel shows the average contribution of the Poisson noise
968:  to the scatter at each phase. Vertical ticks mark the ingress/egress 
969:  phases of the two asymmetric structures seen in the eclipse maps. 
970:  Middle panels: corresponding eclipse maps in a logarithmic grey-scale.
971:  A cross marks the center of the disk and a dotted line shows the 
972:  primary Roche lobe. A solid contour line is overploted on each eclipse
973:  map to indicate the 3-$\sigma$ confidence level region. Right-hand 
974:  panels: asymmetric component of the eclipse maps in a logarithmic 
975:  grey-scale. Additional dotted lines show the ballistic stream 
976:  trajectory, a reference circle of radius $0.6\;R_{L1}$ and semi-circles
977:  of radii 0.2 and $0.32\;R_{L1}$. Labels in the 'ensemble' panel depict
978:  the location of the asymmetric sources 1 and 2. The secondary is to the 
979:  right of each panel and the stars rotate counter-clockwise. Scale bars 
980:  in the right side depict the log(intensity) scale in each case. }
981: \label{fig3}
982: \end{figure*}
983: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
984: 
985: \clearpage
986: 
987: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% FIG.4
988: \begin{figure}
989: \center
990: \includegraphics[scale=0.5,angle=-90]{f4.eps}
991: \caption{Maps of maximum (left) and minimum (right) brightness, obtained 
992:   from the maps of the steady-light and the long-term changes. The 
993:   notation is similar to that of Fig.~\ref{fig3}.}
994: \label{fig4}
995: \end{figure}
996: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
997: 
998: \clearpage
999: 
1000: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% FIG.5
1001: \begin{figure}
1002: \center
1003: \includegraphics[scale=0.5,angle=-90]{f5.eps}
1004: \caption{Average radial intensity distribution of the long-term changes
1005:   map for an assumed distance of 200~pc (BSH). Dotted lines show the 
1006:   1-$\sigma$ limit on the average intensity. Dashed lines depict the 
1007:   radial intensity distributions resulting from the difference between
1008:   steady-state disk models with mass accretion rates differing by 20, 30 
1009:   and 50 per cent. The reference steady-state opaque disk for these 
1010:   calculations has \.{M}$= 10^{-9.1}\,M_\odot\;yr^{-1}$. }
1011: \label{fig5}
1012: \end{figure}
1013: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1014: 
1015: \clearpage
1016: 
1017: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% FIG.6
1018: \begin{figure}
1019: \center
1020: \includegraphics[scale=0.5]{f6.eps}
1021: \caption{Top: the radial run of the relative amplitude of the low- and 
1022:   high-frequency flickering components. Dashed and dotted lines show the
1023:   1-$\sigma$ limits on the average amplitude, respectively for the low-
1024:   and high-frequency flickering. Bottom: the radial run of the fractional
1025:   contribution of the high-frequencies to the total flickering. Dotted 
1026:   lines show the 1-$\sigma$ limits on the distribution. }
1027: \label{fig6}
1028: \end{figure}
1029: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1030: 
1031: %\bibliographystyle{apj}
1032: 
1033: \end{document}
1034: