2 Optical behaviour of AO 0235+16

The long-term monitoring optical light curves of AO 0235+16 in the Johnson's UBV and Cousins' RI bands are shown in Fig. 1.

Figure 1: Long-term monitoring light curves of AO 0235+16 in UBVRI; data are from Kinman & Rieke (1975), Rieke et al. (1976), O'Dell et al. (1978a,b), Pica et al. (1980), Impey et al. (1982), Barbieri et al. (1982), Moles et al. (1985), Smith et al. (1987), Sillanpää et al. (1988a), Webb & Smith (1989), Mead et al. (1990), Sitko & Sitko (1991), Sillanpää et al. (1991), Takalo et al. (1992), Xie et al. (1992), Rabbette et al. (1996), Webb et al. (1997), Takalo et al. (1998), Xie et al. (1999), and Ghosh et al. (2000); data plotted after the vertical line are from the present work.

Figure 2: Light curves of AO 0235+16 in BVRI bands in the years 1996-2000; all data are from the present work.

As generally happens, the best sampled band in the past is the B one, while the use of CCD cameras has led people to observe chiefly in the R band. The data plotted before the vertical line at ${\rm JD}=2450300$ were derived from the literature (see references in the caption); those after are new observations presented in this paper. Thirteen optical observatories all over the world were involved in this project; their names, nationality, position (longitude and latitude), diameter of the telescope and filters used are indicated in Table 1.

 

 

Table 1: List of participating observatories by longitude.

Observatory	Nation	Long. (deg)	Lat. (deg)	Tel. diam. (m)	Filters	Symbola
MSSSO	Australia	+149.066	-31.277	1.0	R	filled triangles
Xinglong	China	+117.575	+40.393	0.6	R	filled squares
Vainu Bappu	India	+78.83	+12.57	2.0	R	asterisks
Abastumani	Georgia	+42.80	+41.80	0.7	R	open circles
Tuorla	Finland	+22.17	+60.27	1.03	V
Roma	Italy	+12.70	+41.80	0.5, 0.7	BVRI	open triangles
Perugia	Italy	+12.372	+43.112	0.4	VRI	open diamonds
Torino	Italy	+7.775	+45.038	1.05	BVRI	open squares
Calar Alto	Spain	-2.546	+37.224	2.2	R	filled circles
Teide	Spain	-16.5	+28.16	0.82	BVRI	crosses
NOT	Spain	-17.88	+28.75	2.56	BVRI	asterisks
Foggy Bottom	USA	-74.25	+43.24	0.4	R	filled stars
Climenhaga	Canada	-123.22	+48.25	0.5	R	filled diamonds

^a The symbols are used to distinguish the R band data plotted in Figs. 3-7.

Seven of them (MSSSO, Xinglong, Vainu Bappu, Calar Alto, Teide, Foggy Bottom, Climenhanga) participated only during the WEBT campaign in November 1997 (see Sect. 2.1) or around that period, while the others provided the long-term optical monitoring. All observers took frames by means of CCD cameras and performed data reduction using standard procedures; the source magnitude was obtained by differential photometry, adopting the reference stars calibrated by Smith et al. (1985) and Fiorucci et al. (1998).

Although the intense source variability makes a comparison among the different datasets difficult to perform, a general agreement was found where the data overlap in time. Many large-amplitude outbursts are visible from Fig. 1; in the B band the minimum and maximum magnitudes observed were $15.09\pm 0.18$ ( ${\rm JD}=2443907$) and $20.5\pm 0.05$ ( ${\rm JD}=2449692$), respectively. In the R band the range of magnitudes spanned is from $14.03 \pm 0.08$ ( ${\rm JD}=2450811$) to $19.47 \pm 0.06$ ( ${\rm JD}=2450278$).

An enlargement of the BVRI light curves in the period 1996-2000 is shown in Fig. 2; an inspection of the R-band curve, the best sampled one, reveals three major peaks, one for each of the last three observing seasons.

Details of the 1997-1998 season are visible in Fig. 3,

Figure 3: Light curve of AO 0235+16 in the Cousins' R band from August 1997 to March 1998; data are from Vainu Bappu (asterisks), Abastumani (open circles), Roma (triangles), Perugia (diamonds), Torino (squares), and Teide (crosses); data from the WEBT campaign occurred from November 1 to 11, 1997 ( ${\rm JD}=2450754$-2450764) are plotted as dark dots.

where different symbols refer to different observatories (see caption to the figure), with the exception of dark dots, which indicate data collected during the WEBT campaign.

The peak at ${\rm JD}=2450811$ was observed by both the Perugia and the Roma Observatories, and corresponding peaks were seen also in the BVI bands (see Figs. 1 and 2). What is quite impressive is the extreme sharpness of the flare: a brightness increase of $1\rm ~mag$ in 4 days (inside a $2.8\rm ~mag$ rise in 27days) was followed by a $1.3\rm ~mag$ decrease in only one day, leading to a $2.0\rm ~mag$ fall in 13 days.

A smoother behaviour characterizes the 1998-1999 observational season (see Fig. 4):

Figure 4: Light curve of AO 0235+16 in the Cousins' R band from July 1998 to January 1999; data are from Abastumani (circles), Roma (triangles), Perugia (diamonds), Torino (squares), Teide (crosses), and NOT (asterisks).

the source was found to be very bright at the beginning of the season, skimming R=14, and then a continuous decrease, interrupted by a couple of flares, led to $R\sim 19$, with a jump of $5\rm ~mag$.

In the last observing season the temporal coverage was definitely worse than in the previous two seasons (see Fig. 2); however, a quite interesting feature is the fall of $1.6\rm ~mag$ in 48 hours occurred in the R band at the beginning of September 1999 ( ${\rm JD}=2451425.657$- 2451427.657).

2.1 The November 1997 WEBT campaign

The WEBT is an international collaboration among astronomers from all the world with the aim of organizing optical campaigns on blazars of specific interest (Mattox 1999a,b; Villata et al. 2000). These campaigns, lasting from a few days to several weeks, are generally undertaken in concert with observations at other wavelengths, and can be triggered by the discovery of a flaring state of an object, usually in the optical band.

This was indeed the case for the first-light WEBT campaign: the November 1997 observations of AO 0235+16 were started after the detection of a considerable brightness increase (about $1\rm ~mag$ in a week) at the end of October 1997 (Webb et al. 1997). An EGRET target of opportunity observation occurred between November 3 and 11, 1997, but only an upper limit of $16 \times 10^{-8} ~ \rm photons \, cm^{-2} \, s^{-1}$ above $100\rm ~MeV$ could be inferred (Hartman, private communication). Observations by RXTE did not find a high X-ray flux (Webb et al. 2000).

The light curve obtained in the first 11 days of November 1997 is plotted in Fig. 5.

Figure 5: Light curve of AO 0235+16 in the Cousins' R band during the first-light WEBT campaign of November 1997; for the explanation of symbols see Table 1.

Data from each observatory are indicated with a different symbol, according to Table 1 (Col. 7). Weather was rather bad in those days in most of western Europe and North America, so that the temporal density of the curve is far from ideal, and far even from the density reached in more recent WEBT campaigns (see Villata et al. 2000, 2001); however, the common observational effort of many observatories around the world led to a satisfactory time coverage.

In order to avoid spurious variations due to possible systematic effects among data from different observatories, in the following only flux changes observed by the same telescope will be considered. A brightness increase of $1.25\rm ~mag$ in two days was detected at the beginning of the campaign and noticeable flux oscillations were observed all the time. Variations of 0.2- $0.4\rm ~mag$ in 5-7 hours were found in a number of cases; very impressive is the dimming of half a magnitude in about 5 hours detected in the night between November 3 and 4. Similar very fast variations have recently been observed by Romero et al. (2000).

Figure 6: A selection of optical spectra from Roma (triangles), Torino (squares), and Teide (crosses); Julian Dates (-2442000) are indicated on the right; straight lines are drawn only to guide the eye through points of the same spectrum; no vertical shift is applied.

2.2 Spectral behaviour in the optical band

Figure 7: The spectral index $\alpha$ as a function of the R flux for the 40 optical spectra from Roma (triangles), Torino (squares), and Teide (crosses).

One item that is interesting to investigate when studying the optical variability of blazars is whether the flux variations are accompanied by spectral changes. In the new data presented in this paper we identified 47 optical spectra collecting data in at least three filters taken by the same telescope, with the requirement that the delay from the first frame to the last one is not greater than 50 min. We discarded 7 spectra which contained large errors and analyzed the remaining 40. A selection of these "well behaved" spectra is presented in Fig. 6 in the $\log \nu$ - $\log (\nu F_{\nu})$ plane, using different symbols for data taken at different telescopes. When taking into account that some discrepancies in the spectral shape can arise from not completely equal photometric systems used in the various observatories, no significant variations can be recognized when the flux changes. However, a deeper inspection reveals that the spread corresponding to the B band is larger than that relative to the I one. This would suggest that the flux variations are of larger amplitude at the higher frequencies, as already observed in other blazars. All 40 "well behaved" optical spectra were then fitted by a classical power law: $F_{\nu} \propto \nu^{\alpha}$ with a $\chi^2$ minimizing procedure. The results are shown in Fig. 7, where the spectral index $\alpha$ is plotted as a function of the flux in the R band and $\alpha$ values with errors greater than 0.25 have been eliminated. One can notice that for high flux levels $\alpha$ presents an almost constant value of -2.8--2.7, while it tends to decrease with decreasing flux, although there is a not negligible spread. This means that the spectrum basically steepens when the source gets fainter, a behaviour that is common to blazars.