Witam wszyscy, Andrzej Kus is sceptical about using superluminal motions for estimating (local) cosmological parameters. Good :).
Andrzej K - please correct me if i'm wrong, but what i understand of your arguments are something like:
-> jets are bent, not systematically, nobody knows why, sometimes helical, very complex
-> superluminal expansion speed is highly variable
-> depends on which blobs are in which place, some blobs are fast, some are slow --> especially in mm data
-> on month scale changes very rapidly - some blobs just stop for some time, and then start up again.
My initial reaction is that in a statistical sample these should average out, provided that the errors are random and symmetrically distributed.
Alternatively: can we try to statistically correct for this?
Given that blobs stop and then start again, or change direction, but are only *detected* when they are moving superluminally, then it would seem to me that the true average speed is slower than estimated from the doppler factor.
Most practical cosmology is about statistical corrections - it requires careful modelling and correct statistical analysis, but it can be done.
Is it reasonable to use the observational data which *shows* the erratic behaviour of these objects to statistically model this?
If yes, then we would have the necessary correction factor.
Anyway, here's a link to Bartek's IAP talk:
http://www2.iap.fr/Conferences/Colloque/col2004/Docs/20040629_Lew.pdf
Andrzej K suggested that Bartek give a presentation on this (a Monday talk?) - seems a fine idea to me, and doesn't require much preparation from Bartek's side (except to take account of Andrzej's comments).
pozdr boud
LISTNAME: cosmo-torun HELP: send an email to sympa@astro.uni.torun.pl with "help" WEB ARCHIVE: http://www.astro.uni.torun.pl/sympa/cosmo-torun/ UNSUBSCRIBE: email to sympa@astro.uni.torun.pl with "unsubscribe cosmo-torun"
witam
On Fri, 27 Aug 2004, Boud Roukema wrote:
Witam wszyscy, Andrzej Kus is sceptical about using superluminal motions for estimating (local) cosmological parameters. Good :).
Andrzej K - please correct me if i'm wrong, but what i understand of your arguments are something like:
-> jets are bent, not systematically, nobody knows why, sometimes helical, very complex
-> superluminal expansion speed is highly variable
-> depends on which blobs are in which place, some blobs are fast, some are slow --> especially in mm data
-> on month scale changes very rapidly - some blobs just stop for some time, and then start up again.
plus it often happens that they fade away very quickly or change shape so within a few months the recognition of the same blob mae sometimes be difficult.
My initial reaction is that in a statistical sample these should average out, provided that the errors are random and symmetrically distributed.
Alternatively: can we try to statistically correct for this?
Given that blobs stop and then start again, or change direction, but are only *detected* when they are moving superluminally, then it would seem to me that the true average speed is slower than estimated from the doppler factor.
hmm this I quite don't understand ? why do you say so, that they are only detected when they are superliuminal. in fact superluminality isn't a condition "sine qua non" for doing this thing.
Most practical cosmology is about statistical corrections - it requires careful modelling and correct statistical analysis, but it can be done.
Is it reasonable to use the observational data which *shows* the erratic behaviour of these objects to statistically model this?
If yes, then we would have the necessary correction factor.
from what I've done and played with the real data, I see that the final value is quite sinsitive on precise measurments from the maps !! so true - the big sample is needed and this should work.
Anyway, here's a link to Bartek's IAP talk:
http://www2.iap.fr/Conferences/Colloque/col2004/Docs/20040629_Lew.pdf
Andrzej K suggested that Bartek give a presentation on this (a Monday talk?) - seems a fine idea to me, and doesn't require much preparation from Bartek's side (except to take account of Andrzej's comments).
that's ok with me. anytime (except for the one, that is needed for repreparation of updated presentation) ;)
bartek
LISTNAME: cosmo-torun HELP: send an email to sympa@astro.uni.torun.pl with "help" WEB ARCHIVE: http://www.astro.uni.torun.pl/sympa/cosmo-torun/ UNSUBSCRIBE: email to sympa@astro.uni.torun.pl with "unsubscribe cosmo-torun"
hi,
On Mon, 30 Aug 2004, Bartosz Lew wrote:
witam
On Fri, 27 Aug 2004, Boud Roukema wrote:
-> on month scale changes very rapidly - some blobs just stop for some time, and then start up again.
plus it often happens that they fade away very quickly or change shape so within a few months the recognition of the same blob mae sometimes be difficult.
OK, but these don't seem to affect the statistics as far as i understand.
It seems to me that blobs that fade are no longer detected, hence no longer relevant; and if a blob changes shape without changing speed, then it does increase the error bar (determining the centre is difficult), or drop out from detection statistics.
My initial reaction is that in a statistical sample these should average out, provided that the errors are random and symmetrically distributed.
Alternatively: can we try to statistically correct for this?
Given that blobs stop and then start again, or change direction, but are only *detected* when they are moving superluminally, then it would seem to me that the true average speed is slower than estimated from the doppler factor.
hmm this I quite don't understand ? why do you say so, that they are only detected when they are superliuminal. in fact superluminality isn't a condition "sine qua non" for doing this thing.
It seems to me that they are more *likely* to be detected when they are superluminal - and the doppler boosting has a big role in this.
Most practical cosmology is about statistical corrections - it requires careful modelling and correct statistical analysis, but it can be done.
Is it reasonable to use the observational data which *shows* the erratic behaviour of these objects to statistically model this?
If yes, then we would have the necessary correction factor.
from what I've done and played with the real data, I see that the final value is quite sinsitive on precise measurments from the maps !! so true - the big sample is needed and this should work.
:)
pozdr boud
LISTNAME: cosmo-torun HELP: send an email to sympa@astro.uni.torun.pl with "help" WEB ARCHIVE: http://www.astro.uni.torun.pl/sympa/cosmo-torun/ UNSUBSCRIBE: email to sympa@astro.uni.torun.pl with "unsubscribe cosmo-torun"
hi
On Mon, 30 Aug 2004, Boud Roukema wrote:
hi,
On Mon, 30 Aug 2004, Bartosz Lew wrote:
witam
On Fri, 27 Aug 2004, Boud Roukema wrote:
-> on month scale changes very rapidly - some blobs just stop for some time, and then start up again.
plus it often happens that they fade away very quickly or change shape so within a few months the recognition of the same blob mae sometimes be difficult.
OK, but these don't seem to affect the statistics as far as i understand.
sure - it's just a problem with measurments.
It seems to me that blobs that fade are no longer detected, hence no longer relevant; and if a blob changes shape without changing speed, then it does increase the error bar (determining the centre is difficult), or drop out from detection statistics.
sure - I agree. but if the blob fades out the whole sequence of previous observations might be useless resulting in having one source less in the sample.
My initial reaction is that in a statistical sample these should average out, provided that the errors are random and symmetrically distributed.
Alternatively: can we try to statistically correct for this?
Given that blobs stop and then start again, or change direction, but are only *detected* when they are moving superluminally, then it would seem to me that the true average speed is slower than estimated from the doppler factor.
hmm this I quite don't understand ? why do you say so, that they are only detected when they are superliuminal. in fact superluminality isn't a condition "sine qua non" for doing this thing.
It seems to me that they are more *likely* to be detected when they are superluminal - and the doppler boosting has a big role in this.
I agree, thats why we try to play only with these guys, but it's a trade of choosing between luminosity (superliminality) and angular resolution ofthe VLBI capabilities measuring proper motions.
Most practical cosmology is about statistical corrections - it requires careful modelling and correct statistical analysis, but it can be done.
Is it reasonable to use the observational data which *shows* the erratic behaviour of these objects to statistically model this?
If yes, then we would have the necessary correction factor.
from what I've done and played with the real data, I see that the final value is quite sinsitive on precise measurments from the maps !! so true - the big sample is needed and this should work.
:)
pozdr boud
----- for previous posts see:
LISTNAME: cosmo-torun HELP: send an email to sympa@astro.uni.torun.pl with "help" WEB ARCHIVE: http://www.astro.uni.torun.pl/sympa/cosmo-torun/ UNSUBSCRIBE: email to sympa@astro.uni.torun.pl with "unsubscribe cosmo-torun"
LISTNAME: cosmo-torun HELP: send an email to sympa@astro.uni.torun.pl with "help" WEB ARCHIVE: http://www.astro.uni.torun.pl/sympa/cosmo-torun/ UNSUBSCRIBE: email to sympa@astro.uni.torun.pl with "unsubscribe cosmo-torun"
On Mon, 30 Aug 2004, Bartosz Lew wrote:
On Mon, 30 Aug 2004, Boud Roukema wrote:
On Mon, 30 Aug 2004, Bartosz Lew wrote:
On Fri, 27 Aug 2004, Boud Roukema wrote:
detected when they are superliuminal. in fact superluminality isn't a condition "sine qua non" for doing this thing.
It seems to me that they are more *likely* to be detected when they are superluminal - and the doppler boosting has a big role in this.
I agree, thats why we try to play only with these guys, but it's a trade of choosing between luminosity (superliminality) and angular resolution ofthe VLBI capabilities measuring proper motions.
Ouch. That's probably something else which needs to be modelled and corrected. Selection effects are what a lot of the hard work is about in observational cosmology... ;)
pozdr boud
LISTNAME: cosmo-torun HELP: send an email to sympa@astro.uni.torun.pl with "help" WEB ARCHIVE: http://www.astro.uni.torun.pl/sympa/cosmo-torun/ UNSUBSCRIBE: email to sympa@astro.uni.torun.pl with "unsubscribe cosmo-torun"
-
Bartosz Lew -
Boud Roukema