Precise and Accurate, or Imprecise and Inaccurate

Eric Flesch writes:
Two new pre-prints with contrasting results are out, kind of like the
scientific equivalent of a food fight. It's a hot topic, the CMB
temperature as a function of redshift -- which, if true, makes any
static model untenable.

[apologies for the much delayed response...]

Add a couple more exhibits:

[1] K. M. Menten, C. L. Carilli, and M. J. Reid (1998),
"Interferometric Observations of Redshifted Molecular Absorption
toward Gravitational Lenses"

"Given that Trot=Tcrnb, the accurate optical depth determinations
afforded by our interferometer measurements allow for a meaningful
estimate of the cosmic microwave background temperature at z = 0.89,
for which big bang theory predicts a value of (1+z)2.73 K = 5.14 K. In
particular, from the HC3N J = 3-2 and 5-4 spectra shown in Fig. 3 we
derive Trot = 4.5+1.5-0.6 K. We note that the errors quoted for the
HC3N rotation temperature are formal uncertainties and do not take
into consideration systematic effects such as variations in the source
covering factor between 14.5 and 24.1 GHz, the frequencies of the
redshifted 3-2 and 5-4 lines."

[2] C. Henkel, K. M. Menten, M. T. Murphy, N. Jethava, V. V. Flambaum,
J. A. Braatz, S. Muller, J. Ott, and R. Q. Mao, A&A 500, 725 (2009)
"The density, the cosmic microwave background, and the
proton-to-electron mass ratio in a cloud at redshift 0.9"

"[...] toward the south-western source, excitation temperatures of
molecular species with optically thin lines and higher rotational
constants are, on average, consistent with the expected temperature of
the cosmic microwave background, TCMB = 5.14 K. However, individually,
there is a surprisingly large scatter which far surpasses expected
uncertainties. A comparison of CS J = 1-0 and 4-3 optical depths
toward the weaker north-western absorption component results in Tex =
11 K and a 1-sigma error of 3 K."

On 24 December arxiv:1212.5456 (accepted by A&A): "A precise and
accurate determination of the cosmic microwave background temperature
at z=0.89" by S. Muller et al determines a CMB temp of 5.08K for PKS
1830-211 at z=0.89, although they stated some assumptions,
particularly page 2 column 2 top "of great importance for our study"
that the emission is behind the absorbing gas.

Note that Muller is a co-author of [2].

On 27 December arxiv 1212.5625 (accepted by ApJ): "On Measuring the
CMB Temperature at Redshift 0.89" by M. Sato et al, determines a CMB
temp of 1.1 - 2.5K for this same galaxy! They pointedly assert that
high-resolution imaging shows that the absorbing gas covers at best
only part of the emitter.

Full author list is Sato, Reid, Menten, Carilli. The latter three
are the authors of [1], and Menten is also a co-author of [2] (with
Muller).

So the 2nd paper clearly refutes the assumption "of great importance"
of the 1st paper, that the absorbing gas covers the whole emitter.
Reading further, the first paper is based on observations done in
2011-2012, the 2nd paper is based on observations done in 1999 (!).

Not making sense? My loose reading is that the 1999 observers found
that the PKS 1830-211 observations did not confirm the CMB temp /
redshift dependency, but did not publish. Now in 2012 a new group of
"young pups" do find such a dependency, in the process ignoring the
findings of the "old timers" of 1999. These old timers are not happy
because their careful observations are being ignored, so are now
publishing their findings at last.

I don't have any personal knowledge, but if I were to make a narrative
to go with the observations, it would go something like this:

(1) Menten et al. make high res observations of PKS 1830-211 and
publish a CMB temp consistent with standard theory, using nominal
assumptions about the systematics

(2) Henkel et al. (including Menten and Muller) observe larger than
expected variability in observations, challenging the assumptions
about systematic effects made in (1)

(3) Muller et al. publish arxiv:1212.5456, essentially claiming that
the systematics found in (2) are well enough understood or small
enough for a precision analysis

(4) Sato et al. publish arxiv:1212.5625, challenging the Muller et
al. claim that the systematics are well understood. I don't really
see this paper as challenging the CMB temp per se, but rather
primarily challenging the systematic uncertainty estimates in Muller
et al.

If you want to really delve into reading the tea leaves, it would seem
that Sato--a relative newcomer to the group--played some significant
role in reanalyzing the 1998 data to test the covering factor
homegeneity.

-dan

On Fri, 01 Mar 13, Dan Riley wrote:
If you want to really delve into reading the tea leaves, it would seem
that Sato--a relative newcomer to the group--played some significant
role in reanalyzing the 1998 data to test the covering factor
homegeneity.

Yes, she is evidently a VLT multi-wave specialist at Max-Planck.
Possibly she will review other papers for similar problems.