1. Long-term variability behaviour of AGN The X Finnish-Russian Radioastronomy Symposium Orilampi, 1-5 September, 2008 Talvikki Hovatta Metsähovi Radio Observatory In collaboration with: M. Tornikoski, A. Lähteenmäki, E. Nieppola, I. Torniainen, M. Lainela, H.J. Lehto, E. Valtaoja, M.F. Aller, H.D. Aller

2. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Outline  Introduction  Sample  Variability timescales Different methodsDifferent methods  Flare characteristics  Conclusions

3. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Introduction  A sample of ~100 AGN has been monitored at Metsähovi Radio Observatory for nearly 30 years  The large dataset enables studies of long- term behaviour  Aim is to better understand the observed properties, radiation mechanisms and physics of the sources

4. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Sample data  22, 37 and 87 GHz data from Metsähovi  4.8, 8 and 14.5 GHz from the University of Michigan (UMRAO)  90 and 230 GHz data from the SEST  90, 150, and 230 GHz data from the literature

5. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Sample sources  80 sources for timescale analysis At least 10 years of monitoring data at 22 or 37 GHzAt least 10 years of monitoring data at 22 or 37 GHz Bright sources with a flux density of at least 1 Jy in the active stateBright sources with a flux density of at least 1 Jy in the active state  55 sources for flare analysis At least one well-monitored distinguishable flare at 2 of the frequencies (22, 37 or 90 GHz being one of the frequencies)At least one well-monitored distinguishable flare at 2 of the frequencies (22, 37 or 90 GHz being one of the frequencies)  Altogether 90 sources  HPQs, LPQs, BLOs and Radio Galaxies

6. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Timescale analyses: methods  Structure function (SF)  Discrete autocorrelation function (DCF)  Lomb-Scargle periodogram  Wavelets Only 22, 37 and 90 GHzOnly 22, 37 and 90 GHz Morelet waveletMorelet wavelet

7. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Timescale analyses: results  Large flares are seen on average every 4 to 6 years Average at 37 GHz is 4.2 years (DCF, wavelets)Average at 37 GHz is 4.2 years (DCF, wavelets)  Redshift-corrected timescales are shorter 2 years for quasars2 years for quasars 3-4 years for BL Lacertae objects (BLOs)3-4 years for BL Lacertae objects (BLOs) -> shocks could be produced less frequently in BLOs  Rise and decay times of flares are between 1 to 2 years

8. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Timescale analyses: results  No strict periodicities were found Episodes of quasi-periodic behaviour are commonEpisodes of quasi-periodic behaviour are common  Multiple timescales are common  Timescales change, get weaker in power or disappear over long time periods

9. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Timescale analyses: results  DCF and wavelets give similar results but wavelets also give information on the continuity of the timescales  Lomb-Scargle periodogram easily produces spurious timescales  Hovatta et al. 2007 (A&A, 469, 899-912), Hovatta et al. 2008b (A&A, in Press)

10. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Example: 4C 29.45 at 22GHz 3.49 years ~3.4 years Flux curve DCF

11. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Example… 3.29 years LS-periodogram 1.21 years SF

12. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Example… 3.4 years 1.7 years

13. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Example: results  Same timescale of ~3.4 years is obtained with all the methods  Only with wavelets it is possible to see that it is present only in the latter half of the flux curve  Comparison of new SF analysis to Lainela & Valtaoja (1993) also showed the difference L&V 1993, SF timescale >6.68 yearsL&V 1993, SF timescale >6.68 years

14. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Flare characteristics  Sample of 55 sources with 159 flares  4.8 – 230 GHz  Flare amplitudes (peak, relative)  Duration of flares  Variability indices  Testing of the shock model  Hovatta et al. 2008 (A&A, 485, 51-61), Nieppola et al. in preparation

15. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Example  Times between flares On average 4 yearsOn average 4 years  Rise / decay times 1-2 years1-2 years  Duration Median 2.5y @ 22 & 37GHzMedian 2.5y @ 22 & 37GHz Range between 0.3-13.2yRange between 0.3-13.2y  Peak flux Median 4.5Jy @22 & 37GHzMedian 4.5Jy @22 & 37GHz Range between 0.7-57 JyRange between 0.7-57 Jy 0.95 years 4 years Peak 4.4Jy Dur 2.5 y

16. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Testing of the shock model α=0.41 α=-0.24 α<-0.5 -Otherwise good general correspondence with the shock model

17. Talvikki Hovatta Metsähovi Radio Observatory The X Finnish-Russian Radio Astronomy Symposium Conclusions  Variability behaviour is complex  Episodes of quasi-periodic behaviour are common  Flares are seen on average every 4 years  Median duration is 2.5 years => Long-term monitoring is essential!