Friday, May 26, 2006

brightening up??

Jackknifes are ready! Had a long discussion this morning with Biebly about what exactly the covariance is and how you calculate it. This was v. informative because now, not being satisfied that I know what IDL is doing, I'm going to modify some of my own code to work out the best-fit line with the covariances taken into account.

Thursday, May 25, 2006

UEFA Co-efficients

But first Luminosity Functions. There have been calls from Shanks to produce a LF for the AAOmega data. Now after having a wee chat with Wake, there are two levels to this. The first is that it is (relatively) easy to perform a 1/Vmax calculation and get a LF. However, this on its own isn't a great scientific achievement. What you want to do is compare this number and see how the LF evolves. This is the second level and comparing AAOmega LRGs to 2SLAQ LRGs could be done but the delta_lookback time is <1 Gyr which in astronomy terms is nothing. Comparing AAOmega LRGs to SDSS LRGs would be sweet but is going to be nigh on impossible.

Meanwhile, I've been figuring out how the hell UEFA work out their seedings in the European and UEFA Cup. Have a wee gander here.

Wednesday, May 24, 2006

I hate it when you run a code for three or four days and then discover you hadn't reset your RR counter and all your xi(s)'s tend to one at large scales. :-( Gutted. Hopefully fixed code and Bielby is running some of the calculations on his machine. Muchos Apprecieatas.

Wolfe (Chicago), of the Sachs-Wolfe affect, gave a enjoyable talk on Damped Lyman-Alpha (DLAs) systems in the Hubble Ultra Deep Field. Shanks is convinced that essentially Lyman-Break Galaxies (LBGs) are today's spirals at redshift~3. Wolfe talked about the (neutral gas?) threshold a galaxy has to be at before star-formation can take place and how all observations so far indicated a low SF density for DLAs (rho_star_dot for DLAs <10^-2.4 M_solar per year per Mpc^3 versus 10^-1.7 - 10^-0.8 M_solar per year per Mpc^3 for LBGs). This very low SF density then throws up problems for metal production ("DLAs are alpha-element enhanced) and heating-cooling rates in these systems.

All very interesting and a bit complicated!

Tomorrow: finish prepping codes for *final* xi(r) jackknifes (still worried bout the IDL linfit routine...)

B.T.W. I think I've found the best radio programme in the world: Drive-Time on talk 107 radio, 4pm-7pm. "Upping the Auntie with Simon Pia & Heather Dee". Superb for a New Town boy.

Tuesday, May 23, 2006

reminding myself

This morning was spent looking at the 120 or so match-ups between GALEX and the 2SLAQ/AAOmega LRGs. Couldn't find any trend of FUV-NUV colour vs. redshift, but there again would you expect to?

This afternoon I reminded myself of the SQL queries I used when doing the S11 COMBO-17 comparisons.
Tomorrow, the jackknifes should be ready!

Monday, May 22, 2006

Hard Rock Hallelujah

Well, the computers finally came back online and got my jackknife program running. Sheesh, I could have done this AT LEAST A YEAR AGO and have to ask myself why I didn't. Anyway, well take all 31 jack-knife samples all the way to xi(r) and then work out the covariance from there...

Had a wee bit of "fun" playing with GALEX data. It's a bit odd that some of the nice GR1 SQL features (such as the link to SDSS DR3) haven't quite been upgraded into GR2. Anyway, will wait until the full AIS data is released.
First look there seems to be ~100 LRGs (2SLAQ and AAOmega) with GALEX FUV and NUV imaging.

Anyway, if this astronomy malarkey doesn't work out there is always, sex, drugs and this.

The Morning Post

So, as has been mentioned in other illustrious blogs (the Mofos), the computers were down over the weekend in the Durham Astronomy department. This has particularly annoyed me since I spent all of Friday modifying my code to run for 48 hours and to pick up the results this a.m. What compounds this situation is that the link to the FORTRAN compiler license is broken so I can't even re-compile my program and get it going again.

And it is blooming miserable outside. Fud.

Friday, May 19, 2006

hee hee hee

Some have commented this blog is a bit "dry". Well, screw you guys.

However I, to quote the Constitution "shall from time to time" deviate from my "research" and post a Challenge on this blog. ALL commers welcome and a small prize (possibly in the shape of a 567 ml glass of alcoholic beverage) shall be offered to the winner*.

So, this weeks challenge is very very simple:

Please send me the best 3 websites in hyper-space**

I shall be looking for: humour factor, geek factor, "Can-I-send-this-to-me-Gran" factor as well
as that mysterious X-factor (WETF that is).

All enteries get "Honourable Mentions" and the Winner gets said pint.
And hey, if you are the only entry, well, de fault are the two most beautiful words in the English language.

Good Night and Good Luck!!

The judge's decision is final. Share prices can go up as well as down.
*If you do not live in the UK, or do not drink, I shall post you £2 equivalent in UK Sterling to your desired address.
**Hyper-space is an abstract idea per se so don't spend too long looking for in it or for it...

Wednesday, May 17, 2006

Eto'o 76, Belletti 80

Made some progress today on the npt code. Fook, this is beginning to be one of "those" weeks but I just don't have many more weeks left.

Conselise (Nottingham) gave a very interesting (though not to say I believed it all) talk this afternoon on morphologies, stellar masses and merger rates of "massive" (10^10-11) galaxies. Bit tired right now (ah....) so will have a think about this some more and report later.

Tuesday, May 16, 2006

another frustrating day

Why, oh why, oh why, oh why?!?!?!?!? Just when I thought it was a simple case of typing "FORMAT" instead of "format" (classic) I then begin to realise that there is something else up and the bloomin' DR counts are still screwy. I'm satisfied my correl_five code is correctly outputting numbers, Wake is satisfied (and I am too) that the npt is fine but what then is happening when I compare the random-random distributions and find the number of DR pairs are less than DD pairs. Oh dear me. Bet The Doctor never has this problem and that's why he gets so much done.

Anyway, "first" thing tomorrow, my own (x,y,z) random points, 1000 and then 10,000 of them and let's see where that leaves us. Sorry if this is turning into a bloan (cross between a blog and a moan).

Some more interesting papers today on the pre-print server.
Cimatti et al. (astro-ph/0605353) commenting on the build-up of mass from z~1 after re-examining rest-frame B-band COMBO-17 and DEEP2 luminosity functions (LFs) of early-type galaxies. From the abstract, "At each redshift there is a critical mass above which virtually all ETGs appear to be in place, and this fits well in the now popular "downsizing" scenario. However, ``downsizing'' does not appear to be limited to star formation, but the concept may have to be extended to the mass assembly itself as the build-up of the most massive galaxies preceeds that of the less massive ones". This seems to fit quite nicely with Wake's recent result and will be v. interesting to hear how it fits in with Conselise (Nottingham) talking tomorrow.

The other paper that caught my eye and that I read was Rigopoulou et al. (astro-ph/0605355) and the discrepancy in the last graph between recent z~3 LBG observations and the ol' theoretical models.

Monday, May 15, 2006

"Eloi, Eloi, lama sabachthani?"

What a frustrating day. Have realised that the "all-singing-all-dancing" npt code doesn't seem to calculate xi(sigma,pi) correctly!! There seems to be an order of magnitude difference between DR pairs which leads to ~unity values of xi and high wp(sigma)'s. ARGH!!! Really not sure what to do. Just when I thought there was only a couple more bridges to cross, this happens!!
If I cannae figure oot what's up with the npt code for the DRs (and RRs????!) then I might just have to resort to what I wanted to do this time last year!! Fook!!!!

On the wires today were Padmanabhan's (Princeton) and Blake's (UBC) photo-z LRG Power spectra papers (astro-ph/0605302 and 0605303). They are basically looking at exactly the same things (angular power spectra with LRGs, photo-z's being estimated from training sets and then some "fancy maths" to get the 3D power spectrum). Well, if correct and they have a handle on all their systematics, photometeric calibrations and other selection issues, then these are (very) powerful pieces of work. However, what I don't understand straight off the bat is how, if you are claiming, at best delta_z = 0.03 on your photo-z, can you then split your sample into bins of 0.05 width (e.g. 0.45 < z <0.50) ????

Sunday, May 14, 2006

That goal was perfection

Oh dear. It is grey and raining outside on a fairly relaxing and very quiet Sunday afternoon.

Thursday and Friday were spent continuing working on xi(sigma,pi) refinements and lo and behold, Shanks read my paper and made some good useful comments. After finishing off a couple of things on Monday morning, it'll be full head steam ahead as I feel we really are close but just have to take a deep breath and finish it off.

Have just been checking out various astronomy websites (Spitzer, GMT, ESO) but have very little of any value to report. Sorry.

Wednesday, May 10, 2006

understanding perldl

Hmmm. Not my most productive day ever. Spent the morning going back over the .pl scripts I have to calculate covariance matrices and correlation functions to make sure all the errors tallied to the right bins. They did and I do essentially now have (for the first ever time) a xi(s) for the entire 2SLAQ LRG Sample 8 with proper jackknife errors. Bout ruddy time too.

Lunchtime talk was given by the IPPP and the statement to discuss was "WMAP3 rules out inflation models with potentials greater than or equal to phi^4." Hmmmm, interesting but didn't quite understand (like many of the audience) what assumptions go into making this claim. Also "racetrack" potentials are still favoured. However, it does seem as if n_s is red...

Spent most of the afternoon reading-up on basic astrophysics and jotting down a few things for my "Employmen Information Package".

Shanks in Nottingham today. Nuff said.

Tomorrow am: Bulk up jots and re-jig Statement of Research Interests.
Tomorrow pm: FITS files and sodding xi(sigma, pi) (leftover from today!)

Tuesday, May 09, 2006

Serbia and Montenegro vs. Ivory Coast

I officially published the AAOmega LRG Pilot Catalogue today. Well, at least sent a file with 1360 objects observed by AAOmega with positions, redshifts, "qops" and magnitudes galore to those in Australia that have been requesting said data. Nice to have been in charge of that from conception to publication. Now to write it up properly!! Bielby is working on the ELGs from the Pilot run and we should have those by the end of the week.

3 really nice contributions to "Observers' Lunch" today. Alexander (IoA -> Durham) talking about Chandra 2Ms exposure, obscured AGN, dusty tori and Spitzer galaxies. Mullaney giving us a first glimpse into what some of his first year work has been on (AGN spectra modelling from radio to X-rays) and Schurch reviewing the discovery announced at NAM2006 of a new type of pulsar which no-one has any idea about. Cool.

Still waiting for feedback from Shanks....

Tomorrow: 1) xi(sigma,pi) .fits files and 30 calculations of wp(sigma) 2) Employment Information Package!!

Monday, May 08, 2006

From Paris to Berlin

....and every disco I get in.

Produced a nice graph of Poisson errors vs. Jackknifes today. Very comparable on "small" scales (<5 Mpc) but then the Poisson errors (badly) underestimate the spread on the larger scales and my hump of 3 points at ~40 Mpc have the largest errorbars. Superb.

Been looking at the COMBO-17 galaxies we got redshifts for. Looks like there could be something here as the COMBO photoz's seem to be systematically smaller than the spectroscopic AAOmega qop 3's.

Shanks did not read my paper over the weekend. Maybe tomorrow...

Sunday, May 07, 2006

Sunday evening blues

Glorious Friday was spent running the npt code for me deprojected correlation function jackknifes. Left the RR case running over the weekend.

Lunchtime talks were by Angulo and Helly, who talked about BAO predictions from semi-analytic models ("ICC1340") and (lack of evidence for) cirlces in the CMB and a non-infinite Universe. Anyway, nice to know that we live in a Universe which is at least 26 Gpc big (what that means exactly I'm not entirely sure).

Gave the latest draft of the paper to Shanks. Let's see if he reads it over the weekend....

Thursday, May 04, 2006

Fortuna Imperatrix Mundi, Fortune Plango Vulnera Una

I did some maths today. Begun to get my head around covariance and correlation matrices and that Numerical Receipes book is a good as anything. Feel I understand now where all the covariances link in, at least as far as the "Normal Formulae" go.

Also - and I apologies to those that think I am a heathen for only doing this now - had my first good experience today with IDL. Thanks Norris and da Angela (both Durham). As such, have a nice least squares function waiting for me tomorrow!

However, late this evening figured out why I've been so worried about this whole thing. Essentially I'm still only working out variances and covariances on the redshift-space correlation function (de-projected or otherwise). However, you still have to collapse/integrate this to get wp(sigma) and then INTERPOLATE to get the real-space correlation function. Argh! And how the errors carry through is anyones "guess".

bruised feet

Went back to editing/tidying-up and writing the redshift-space distortion paper today. In the process of do so and re-doing an N(z) histogram, discovered that I could improve the random catalogue redshift distribution so it matched the LRG data much more closely. Achieved this and re-calculated (once a friggin' 'gain) the correlation function. Have stated to look into the errors again and how to manipulate the covariance matrix but that can be Thursday job.

Tuesday, May 02, 2006

hiding behind my jumper

More work on the thesis today, which consisted of re-writing the ol' paper. Realised that the logic in the Results section is a bit screwy so had a go at re-ordering this into a sensible progression.

Not much more news. Wasn't a slow day as such, just really "erg-ging" so tae speak. Had an IPPP talk at lunch which was kinda interesting again about WMAP3 but had heard most of the stuff before. Still always worth going to these things as you do always tend to learn or reinforce something.

Monday, May 01, 2006

And so it begins....

I officially started writing my thesis today.

As a result, today was a grind sorting out my refs.bib file (i.e. the Bibliography from NASA ADS).

I would have been ~2 hours more productive, if just after lunch I hadn't deleted said file by mistake. Dam it.