Wednesday, March 30, 2005

xmmpsf model

I suspect that the xmmpsf model in v12 could be speeded up. The xmmFact method copies the PSF array inside the target region loop which cannot be the most efficient approach.

Keywords: xspec, heasoft

CCFits fitsPointer() method

This method appears to return the fitsfile* for the last HDU in the file regardless of the current HDU. To correct this set the HDUposition element of the fitsfile struct. This should be fixed in CCfits at some point.

Keywords: heasoft, CCfits

Saturday, March 26, 2005

Extra models

Added the URL for the extra models page to the end of the list of available models in xspec (both v11 and v12). If we move to a browser help in v12 may be able to automatically add these extra models to the on-line help.

Keywords: xspec, heasoft

Opt/UV extinction model

Martin Still extended the Cardelli et al. extinction routine down to 909 Angstroms . Added a redshifted version - required for Swift UVOT data.

Added the model to both v11 and v12 although still need to modify documentation for v12.

Keywords: xspec, heasoft

Wednesday, March 23, 2005

xspec v12 local model issues

Tried out building v12 local models. Firstly, note that initpackage didn't run standalone because it looks for environment variable XSPECROOT which is set when xspec starts up. Fixed this by looking instead for HEADAS. Also, trapped case of HEADAS not being set (which means getenv returns a null pointer). Note that when running standalone the user has to follow initpackage with an hmake in the directory with the local model code. When using the initpackage command from within xspec this happens automatically.

Old v11 models that include don't work at the moment. initpackage needs to copy the
include/xspec.h file into an file in the directory with the local model code.

The v12 local model paradigm doesn't work under OS X for fortran routines including common blocks because these cannot be placed in a dynamic library. Note that any fortran routine which includes falls in this category. We will have to either require no fortran common blocks in local models under OS X or provide an alternative mechanism to build static libraries which would be linked into the xspec executable when it is built.

Keywords: xspec, heasoft

Tuesday, March 22, 2005

xspec 11.3.2

Updated the version number for xspec 11 to 11.3.2 for the new release. This is essentially just a bug fix release with a couple of the extra models now included automatically.

Keywords: xspec, heasoft

Friday, March 18, 2005


Pawl et al. create mock Astro-E2 observations from their simulated clusters.

Seager et al. consider the spectroscopic biosignature of extrasolar planets.

Gies & Hesel claim that ice ages correspond to the Sun traversing spiral arms.

Sanders et al. present deep Chandra imaging of the Perseus cluster. They find a high-abundance shell which they suggest is the edge of a fossil bubble. They also detect non-thermal emission with a total luminosity of 4.8x10^43 erg/s and use this to map the magnetic field in the core of the cluster.

Weisskopf et al. provide a detailed description of the Chandra hardware and in-orbit performance.

de Diego et al. argue that the Pioneer anomalous acceleration is due to the Kuiper belt.

Keywords: clusters, extrasolar planets, Chandra, Pioneer

Katherine Blundell: SS433

Seminar by Katherine Blundell of Oxford on optical and radio observations of microquasars and in particular SS433. Neatest result was analysis of new VLA image. She showed that assuming SS433 shoots ballistic plasma blobs out completely symmetrically then the differences in observed shape from the two corkscrewing jets are due to relativistic aberration. This enables the jet speed to be determined accurately to 0.26c and hence the distance to 5.5 kpc. She also noted that the model is not a perfect fit and can be improved by allowing the speed to vary between 0.24 and 0.28c. Further she showed that by combining the redshifts of symmetric blobs one can determine independently the velocity and the cone opening angle. A PSD of the velocity shows a signal at the binary frequency. This may be due to eccentricity in the orbit.

Keywords: microquasars, SS433

xspec 11 seg fault in fakeit

Bryan pointed out a seg fault under OSF in xspec v11 when running a fakeit. Turned out to be an attempt to strcpy a null string. Strange that it hadn't turned up before. Will be fixed in the forthcoming heasoft release.

Keywords: xspec, heasoft

Monday, March 14, 2005


Fabian et al. present initial results from a 200 ksec Chandra observation of the Centaurus cluster. They see bubbles, semi-circular edges and filaments. Some edges have sharp temperature increases and abundance drops.

Barcons et al. find a single absorption line in a 93 ksec XMM-Newton RGS spectrum of the Bl Lac PKS 0548-322. If this line is OVI then the absorber is at z=0.058. The column is however much larger than expected.

Rhook & Wyithe calculate LISA detection rates for SMBH mergers and predict ~15/year with S/N > 5.

Bauer et al. use Chandra to study 38 X-ray luminous clusters in z ~0.15-0.4 and find 55% have cool cores and 34% have t_cool less than 2 Gyr.

Zane et al. find a possible absorption feature in the XMM PN&MOS spectra of the isolated NS RBS 1774.

Keywords: clusters, IGM, SMBH, GW, NS

Downsampling in HHT

I don't see how this can work in the way suggested. In any IMF Omega varies smoothly on timescales longer than the averaging time (as it must because faster variation has been removed in earlier IMFs). We do not gain a sqrt(N) advantage which would require Omega to vary randomly over an averaging time.

Keywords: HHT, GW

Hilbert transform in HHT code

Examining Omega vs Time in individual IMFs showed a high frequency oscillation - there is a clear trend with successive bins above and below the trend. This oscillation does not show up in Delta but does in Kappa (the Hilbert transform). This led me to examine the Hilbert transform code. The Hilbert transform algorithm is to forward FT the time series, modify the result, then inverse FT. The imaginary part of the result is the Hilbert transform while the real part should be the input time series. Comparing the input time series with the real part of the result showed a difference at the ~1% level. This is much more than the numerical accuracy of the code.

Googling to find other descriptions of the algorithm I found a minor error in the code. The first and Npoints/2 + 1 points in the FFT should be left unchanged - not multiplied by 2
(like the 2 - Npoints/2 points). Fixing this removed the difference between the input and output time series. Unfortunately, it still left the oscillation in the transform.

I don't understand the origin of this oscillation at present but I was able to remove it by running a three-point triangular filter over the transform array. This in turn removed the high-frequency oscillation in the Omega vs Time plot.

Keywords: HHT, gravitational waves

Saturday, March 05, 2005

error in Huang et al. 1998

Equations 6.3 and 6.4 in Huang et al. Proc.R.Soc.Lond. A, 454, 903 are in error. In the double summation j should run from k+1 to n+1. I suspect there should also be a 2 in front of the summation. Note of course that X^2(t) should be outside the parentheses.

Keywords: HHT, GW


Added Juri Poutanen's compps model (version 5/12/04) to xspec. Modified the comprefl.f and xscompps.f to call greencj from the pexriv model instead of greeion included in comprefl. This fixes a bug under Linux and removes replicated source code. This could lead to problems if pexriv and compps are run simultaneously although there is little physical justification for that.
Need a couple of paragraphs from Juri for the help.

Keywords: xspec, heasoft, comptonization

Thursday, March 03, 2005


Mullis et al. have serendipitously discovered an X-ray luminous cluster at z=1.4. The velocity dispersion is 760+/-260 and there are a large number of red galaxies.

Keywords: clusters

Michael Mumma: Methane on Mars

Mike Mumma gave the Goddard Science Colloquium last Friday and described recent work detecting methane on Mars. This has a rocky history as there have been several claims over the last 20-30 years which have not been confirmed. The significance of a detection is that most methane on Earth is biotic. Methane on Mars could be abiotic but that would imply that the planet is geologically active which is not thought true at the moment. Mumma emphasized the importance of simultaneous imaging because ice clouds in the Martian atmosphere will reduce the column observed hence change the calculated abundance. His team have been performing long-slit near-IR spectroscopy using IRTF and Gemini South and they have clear detections of methane which is spatially variable. They are now looking for correlations with ground features, temperature etc. Once the geographic sources of methane have been identified then other tests can be performed, ideally involving in situ measurements of 12C/13C ratios which are diagnostics of biotic origins.

Keywords: astrobiology, Mars

Wednesday, March 02, 2005

Biman Nath: Entropy in the ICM and AGN

Biman Nath gave the UMCP Astronomy Colloquium. He reviewed the entropy of the ICM and in particular that in groups. S = T/n^2/3. If clusters are created at the same time we expect S to go as T but it goes as T^0.65. Further we expect L to go as T^2 however it goes as almost T^3. These imply something other than gravity determines the structure of the ICM. The effect is particularly marked in groups. One possibility is that galaxies form more efficiently in groups leading to the low S gas to cool out and be replaced by high S gas falling into the center. This may have some effect but cannot explain the observations on its own. The alternative is heating either by AGN or cosmic rays. There is good evidence now for AGN energy input into the ICM. The unknown is whether this can happen at higher redshifts and to the extent required. A fascinating new result (Croston et al 2005) seems to show that those groups containing radio-loud AGN have higher entropy than those having radio-quiet AGN. I pointed out in a question that since entropy is a time integral over the lifetime of the cluster this would imply that AGN which are radio-loud now were radio-loud in the past (and similarly for radio-quiet). Nath wondered whether the observed entropy would change while a radio-loud AGN was active.

Keywords: clusters, AGN, entropy

Tuesday, March 01, 2005


Cooray & Seto argue that LISA will be able to determine the distance to the LMC using observations of close WD binaries. Their estimated uncertainty is competitive to that from SIM or GAIA.

Merritt et al. argue that luminosity distributions of early-type galaxies and mass profiles of LCDM halos can be described by a Sersic law (log slope of projected density is a power-law of radius). This works better than NFW for LCDM halos.

Enoki et al. calculate gravitation radiation from merging SMBHs using a semi-analytic model of galaxy and quasar formation.

Mainieri et al. look at optically faint X-ray sources in the CDFS and use photometric redshifts to examine the distribution of sources too faint for spectroscopy. They still find an overdensity of z<1 sources wrt background synthesis models. 25% of the sources are absorbed QSOs.

Anania & Makoid argue that the Pioneer 10 anomalous acceleration can be explained by bending of background gravity behind the Sun (analogous to light bending).

Keywords: Gravitational waves, dark matter halos, AGN

error command in v12

I went through the XSfit/Fit/FitErrorCalc.cxx code and made changes to match those in the v11 error command added over the last year. The changes included the recent fix for when the soft and hard limits differ (clcunc.f 1.27), the fix for numerical accuracy problems near the hard limits (clcunc.f 1.26), setting the initial trial to a maximum of half the hard limit if the default method gives a value outside the hard limit (clcunc.f 1.23), and the randomized choice of a value between xmin and xmax instead of always picking (xmin+xmax)/2 (clcunc.f 1.21).

Questions still to consider: does v12 handle eigenvalue data correctly (clcunc.f 1.24) ? Are new parameters saved correctly if a new minimum is found (clcunc.f 1.25) ? Should the random number generator be initialized (clcunc.f 1.22) ? What happens if curdel is < ylimit(1) (clcunc.f 1.22) ?

UPDATE: Included the clcunc.f 1.22 changes. The others require discussion with Craig.

Keywords: heasoft, xspec