Planck lensing likelihood corrections (1502.01591)

Use of Cobaya. camb, CLASS, cosmomc, compilers, etc.
Post Reply
Mathew Madhavacheril
Posts: 3
Joined: February 01 2015
Affiliation: Princeton University
Contact:

Planck lensing likelihood corrections (arXiv:1502.01591)

Post by Mathew Madhavacheril » November 24 2015

Hi,

I'm trying to verify the statement that the corrections to the Planck lensing likelihood (detailed in C.5 of Planck Lensing 2015 1502.01591) should be zero when the input cosmology is equal to the fiducial cosmology used in the lensing analysis.

To do this, I use the "test likelihood" function in CosmoMC and enable only "lensing.ini". I then set the cosmological parameters to the fiducials specifed in 1502.01591 . I run the likelihood test and it tells me Planck lensing provides a loglike of 5.3.

Next, I go into source/CMBLikes.f90 and comment out the correction by changing, inside GetBinnedMapCls,

Code: Select all

 Cls = Cls + (correctionCl - this%FiducialCorrection(:,bin))
to

Code: Select all

 Cls = Cls 
because that's what I thought would mess with the rest of the code (and data) the least.

I run the CosmoMC test again and it tells me Planck lensing provides a loglike of 4.67 .

So my question is, shouldn't these two loglikes be exactly equal, if not very close?

The input parameters I'm using are:

Code: Select all

param[omegabh2] = 0.0222
param[omegach2] = 0.1203
param[omeganh2] = 0.00064
param[theta] = 1.041281
param[tau] = 0.065

num_massive_neutrinos=1
param[mnu] = 0.06
param[meffsterile] = 0

param[omegak] = 0
param[w] = -1
param[nrun] = 0
param[nrunrun] = 0
param[r] = 0

param[wa] = 0
param[nnu] = 3.046
param[yhe] = 0.24

param[alpha1] = 0
param[deltazrei] = 0.5
param[Alens] = 1
param[Alensf]=-1
param[fdm] = 0

param[ns] = 0.96
#log[10^10 A_s]                                                                                         
param[logA] = 3.04

param[Aphiphi] = 1

param[calPlanck]= 1

Note that since 1502.01591 specifies H0 instead of theta, I used CAMB to convert \{others,H0\} to \{others,theta\}. Also I've set param[calPlanck]= 1 .

Thanks,
Mat

Antony Lewis
Posts: 1941
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

Re: Planck lensing likelihood corrections (arXiv:1502.01591)

Post by Antony Lewis » November 24 2015

Could you set test_output_root = test-ouputput.txt so we can compare the parameters it is actually calculating? (For example the Helium number may not be consistent.)

The CAMB input file corresponding to FFP8 and the lensing fiducial is

FFP8_params.ini

though note that T/E were rescaled to matched the actual data power as described in the 2015 lensing paper.

Mathew Madhavacheril
Posts: 3
Joined: February 01 2015
Affiliation: Princeton University
Contact:

Planck lensing likelihood corrections (arXiv:1502.01591)

Post by Mathew Madhavacheril » November 24 2015

I've tried to make the CosmoMC parameter inputs and the CAMB inputs that convert H0 to theta as consistent as possible with the FFP8 parameters. The output .par file from CosmoMC now reads:

Code: Select all

 -log(Like) =    5.18516669523127
  chi-sq    =    10.3703333904625


    1  0.2218376E-01   omegabh2              \Omega_b h^2
    2  0.1202957E+00   omegach2              \Omega_c h^2
    3  0.1041180E+01   theta                 100\theta_{MC}
    4  0.6504258E-01   tau                   \tau
    5  0.0000000E+00   omegak                \Omega_K
    6  0.6000000E-01   mnu                   \Sigma m_\nu
    7  0.0000000E+00   meffsterile           m_{\nu,{\rm{sterile}}}^{\rm{eff}}
    8 -0.1000000E+01   w                     w
    9  0.0000000E+00   wa                    w_a
   10  0.3046000E+01   nnu                   N_{eff}
   11  0.2448930E+00   yhe                   Y_{He}
   12  0.0000000E+00   alpha1                \alpha_{-1}
   13  0.5000000E+00   deltazrei             {\Delta}z_{\rm re}
   14  0.1000000E+01   Alens                 A_{L}
   15 -0.1000000E+01   Alensf                A^f_L
   16  0.0000000E+00   fdm                   \epsilon_0 f_d
   17  0.3040035E+01   logA                  {\rm{ln}}(10^{10} A_s)
   18  0.9595744E+00   ns                    n_s
   19  0.0000000E+00   nrun                  n_{\rm run}
   20  0.0000000E+00   nrunrun               n_{\rm run, run}
   21  0.0000000E+00   r                     r
   22  0.0000000E+00   nt                    n_t
   23  0.0000000E+00   ntrun                 n_{t,{\rm run}}
   24  0.1000000E+01   Aphiphi               A^{\phi\phi}_L
   25  0.1000000E+01   calPlanck             y_{\rm cal}

   26  0.6718201E+02   H0                    H_0
   27  0.6828912E+00   omegal                \Omega_\Lambda
   28  0.3171088E+00   omegam                \Omega_m
   29  0.1431246E+00   omegamh2              \Omega_m h^2
   30  0.6451439E-03   omeganuh2             \Omega_\nu h^2
   31  0.9615398E-01   omegamh3              \Omega_m h^3
   32  0.8098034E+00   sigma8                \sigma_8
   33  0.4560199E+00   s8omegamp5            \sigma_8 \Omega_m^{0.5}
   34  0.6076894E+00   s8omegamp25           \sigma_8 \Omega_m^{0.25}
   35  0.9879912E+00   s8h5                  \sigma_8/h^{0.5}
   36  0.2454246E+01   rmsdeflect            \langle d^2\rangle^{1/2}
   37  0.8796692E+01   zrei                  z_{\rm re}
   38  0.2090598E+01   A                     10^9 A_s
   39  0.1835588E+01   clamp                 10^9 A_s e^{-2\tau}
   40  0.1215541E+04   DL40                  D_{40}
   41  0.5595436E+04   DL220                 D_{220}
   42  0.2469914E+04   DL810                 D_{810}
   43  0.7919257E+03   DL1420                D_{1420}
   44  0.2230794E+03   DL2000                D_{2000}
   45  0.9595744E+00   ns02                  n_{s,0.002}
   46  0.2453093E+00   yheused               Y_P
   47  0.2466355E+00   YpBBN                 Y_P^{\rm{BBN}}
   48  0.2626574E+01   DHBBN                 10^5D/H
   49  0.1380986E+02   age                   {\rm{Age}}/{\rm{Gyr}}
   50  0.1090181E+04   zstar                 z_*
   51  0.1444972E+03   rstar                 r_*
   52  0.1041379E+01   thetastar             100\theta_*
   53  0.1387557E+02   DAstar                D_{\rm{A}}/{\rm{Gpc}}
   54  0.1059513E+04   zdrag                 z_{\rm{drag}}
   55  0.1472220E+03   rdrag                 r_{\rm{drag}}
   56  0.1405862E+00   kd                    k_{\rm D}
   57  0.1610484E+00   thetad                100\theta_{\rm{D}}
   58  0.3404822E+04   zeq                   z_{\rm{eq}}
   59  0.1039186E-01   keq                   k_{\rm{eq}}
   60  0.8125179E+00   thetaeq               100\theta_{\rm{eq}}
   61  0.4491191E+00   thetarseq             100\theta_{\rm{s,eq}}
   62  0.7130328E-01   rsDv057               r_{\rm{drag}}/D_V(0.57)
   63  0.9286088E+02   Hubble057             H(0.57)
   64  0.1393041E+04   DA057                 D_A(0.57)
   65  0.6774476E+00   FAP057                F_{\rm AP}(0.57)
   66  0.4722252E+00   fsigma8z057           f\sigma_8(0.57)
   67  0.6010880E+00   sigma8z057            \sigma_8(0.57)
   68  0.1037033E+02   chi2_lensing          \chi^2_{\rm lensing}
   69  0.0000000E+00   chi2_prior            \chi^2_{\rm prior}

 -log(Like)     chi-sq   data
      5.185     10.370   CMB: lensing = smica_g30_ftl_full_pp

The agreement with and without corrections now improves slightly (5.19 vs 4.69).

H0 still comes out slightly off. Here is the CAMB input from which I get theta for the FFP8 H0:

Code: Select all

output_root = planck_lensing_2015
get_scalar_cls = T
get_vector_cls = F
get_tensor_cls = F
CMB_outputscale = 7.42835025e12
get_transfer = F
accuracy_boost = 1
l_accuracy_boost = 1
high_accuracy_default = T
do_nonlinear = 0
l_max_scalar = 4000
k_eta_max_scalar = 8000
do_lensing = T
lensing_method = 1
w = -1
cs2_lam = 1
hubble = 67.11670
use_physical = T
ombh2 = 0.02218376
omch2 = 0.1202957
omnuh2 = 0.00064
omk = 0
temp_cmb = 2.7255
helium_fraction = 0.244892959523355
massless_neutrinos = 2.046
nu_mass_eigenstates = 1
massive_neutrinos = 1
share_delta_neff = T
nu_mass_fractions = 1
DebugParam =   0.000000000000000E+000
Alens =    1.00000000000000
reionization = T
re_use_optical_depth = T
re_optical_depth = 0.06504258
re_delta_redshift = 0.5
re_ionization_frac = -1
pivot_scalar = 0.05
pivot_tensor = 0.05
initial_power_num = 1
scalar_spectral_index(1) = 0.9595744
scalar_nrun(1) = 0
scalar_nrunrun(1) = 0
scalar_amp(1) = 2.090598E-09
RECFAST_fudge_He = 0.86
RECFAST_Heswitch = 6
RECFAST_Hswitch = T
RECFAST_fudge = 1.14
AGauss1 =  -0.140000000000000
AGauss2 =   7.900000000000000E-002
zGauss1 =    7.28000000000000
zGauss2 =    6.73000000000000
wGauss1 =   0.180000000000000
wGauss2 =   0.330000000000000
do_lensing_bispectrum = F
do_primordial_bispectrum = F
initial_condition = 1
scalar_output_file = scalCls.dat
lensed_output_file = lensedCls.dat
lens_potential_output_file = lenspotentialCls.dat
accurate_polarization = T
accurate_reionization = T
accurate_BB = F
derived_parameters = T
version_check = Feb15
do_late_rad_truncation = T
feedback_level = 1
massive_nu_approx = 1
number_of_threads = 0
use_spline_template =  T
l_sample_boost = 1
What do you think is causing CosmoMC to get a derived H0 that is different from what I input to CAMB?

Antony Lewis
Posts: 1941
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

Re: Planck lensing likelihood corrections (arXiv:1502.01591)

Post by Antony Lewis » November 25 2015

yheused looks different, probably due to the updated BBN interpolation table for the latest CosmoMC. Try turning off BBN consistency in Cosmomc so YHe is set explicitly?

Post Reply